id stringlengths 33 34 | doc_url stringlengths 33 34 | title stringlengths 10 211 | publication_date timestamp[s] | update_date timestamp[s] | authors stringlengths 5 20.1k | category_primary stringclasses 78
values | category_all stringlengths 5 138 | abstract stringlengths 123 3.09k | embeddings listlengths 768 768 |
|---|---|---|---|---|---|---|---|---|---|
http://arxiv.org/abs/2406.13348v1 | http://arxiv.org/pdf/2406.13348v1 | Textual Unlearning Gives a False Sense of Unlearning | 2024-06-19T08:51:54 | 2024-06-19T08:51:54 | Jiacheng Du, Zhibo Wang, Kui Ren | cs.CR | cs.CR, cs.AI, cs.CL, cs.LG | Language models (LMs) are susceptible to "memorizing" training data,
including a large amount of private or copyright-protected content. To
safeguard the right to be forgotten (RTBF), machine unlearning has emerged as a
promising method for LMs to efficiently "forget" sensitive training content and
mitigate knowledge l... | [
0.1308786273,
-0.8585812449,
0.1991975904,
-0.4531799555,
-0.3121695518,
0.3653580546,
0.083030507,
-0.1873614639,
-0.5214146972,
0.8781451583,
-0.4952700436,
-0.1740530431,
-0.642680943,
-0.5104648471,
-0.6891733408,
1.078771472,
0.2186450213,
-0.0762083456,
-0.2376409471,
-0.... |
http://arxiv.org/abs/2406.13342v1 | http://arxiv.org/pdf/2406.13342v1 | ZeroDL: Zero-shot Distribution Learning for Text Clustering via Large Language Models | 2024-06-19T08:48:05 | 2024-06-19T08:48:05 | Hwiyeol Jo, Hyunwoo Lee, Taiwoo Park | cs.CL | cs.CL, cs.AI | The recent advancements in large language models (LLMs) have brought
significant progress in solving NLP tasks. Notably, in-context learning (ICL)
is the key enabling mechanism for LLMs to understand specific tasks and
grasping nuances. In this paper, we propose a simple yet effective method to
contextualize a task tow... | [
-0.0323152132,
-0.8294947743,
0.7605637908,
-0.0188273322,
-0.5211240053,
-0.1137155294,
-0.3512728214,
-0.0413671322,
0.1021785364,
0.3120615184,
-0.4612133503,
-1.159825325,
-0.449395895,
-0.0904626027,
-0.664441824,
1.1808637381,
0.0829534158,
-0.199425295,
-0.2829062045,
-0... |
http://arxiv.org/abs/2406.13340v1 | http://arxiv.org/pdf/2406.13340v1 | SD-Eval: A Benchmark Dataset for Spoken Dialogue Understanding Beyond Words | 2024-06-19T08:46:29 | 2024-06-19T08:46:29 | Junyi Ao, Yuancheng Wang, Xiaohai Tian, Dekun Chen, Jun Zhang, Lu Lu, Yuxuan Wang, Haizhou Li, Zhizheng Wu | cs.CL | cs.CL, cs.SD, eess.AS | Speech encompasses a wealth of information, including but not limited to
content, paralinguistic, and environmental information. This comprehensive
nature of speech significantly impacts communication and is crucial for
human-computer interaction. Chat-Oriented Large Language Models (LLMs), known
for their general-purp... | [
-0.3827030957,
-0.6896551847,
0.678429842,
0.125969559,
-0.1852441877,
0.1913171858,
-0.4014739096,
-0.112749815,
-0.1384409219,
0.6114341021,
-0.66513592,
-0.6554560661,
-0.3076552153,
-0.1405952573,
-0.3506334722,
0.9698012471,
0.2240156531,
0.2481651306,
-0.3686770499,
-0.26... |
http://arxiv.org/abs/2406.13337v1 | http://arxiv.org/pdf/2406.13337v1 | Medical Spoken Named Entity Recognition | 2024-06-19T08:39:09 | 2024-06-19T08:39:09 | Khai Le-Duc | eess.AS | eess.AS, cs.CL, cs.LG, cs.SD | Spoken Named Entity Recognition (NER) aims to extracting named entities from
speech and categorizing them into types like person, location, organization,
etc. In this work, we present VietMed-NER - the first spoken NER dataset in the
medical domain. To our best knowledge, our real-world dataset is the largest
spoken NE... | [
-0.1449896991,
-0.5046858788,
0.5493131876,
-0.1663215011,
-0.2330766469,
0.0367902517,
-0.2262344062,
-0.5277629495,
0.4524050951,
0.8994941115,
-0.0851149336,
-0.6869450808,
-0.9080202579,
0.5096852779,
-0.5099067688,
0.9298850298,
0.1895982772,
0.5506267548,
-0.0471538492,
-... |
http://arxiv.org/abs/2406.13332v1 | http://arxiv.org/pdf/2406.13332v1 | How effective is Multi-source pivoting for Translation of Low Resource Indian Languages? | 2024-06-19T08:31:52 | 2024-06-19T08:31:52 | Pranav Gaikwad, Meet Doshi, Raj Dabre, Pushpak Bhattacharyya | cs.CL | cs.CL | Machine Translation (MT) between linguistically dissimilar languages is
challenging, especially due to the scarcity of parallel corpora. Prior works
suggest that pivoting through a high-resource language can help translation
into a related low-resource language. However, existing works tend to discard
the source senten... | [
-0.2153874934,
-0.4880403876,
0.4551415741,
0.3010503352,
-0.6224439144,
0.3370029628,
-0.5309761763,
-0.2218891233,
0.1586726755,
0.2417259961,
-0.4763554037,
-0.0006580298,
-0.8224565387,
0.4437360466,
-0.5558258891,
0.7412512898,
-0.0714788809,
0.256578356,
0.1698225141,
-0.... |
http://arxiv.org/abs/2406.13331v1 | http://arxiv.org/pdf/2406.13331v1 | Improving Zero-shot LLM Re-Ranker with Risk Minimization | 2024-06-19T08:29:54 | 2024-06-19T08:29:54 | Xiaowei Yuan, Zhao Yang, Yequan Wang, Jun Zhao, Kang Liu | cs.CL | cs.CL | In the Retrieval-Augmented Generation (RAG) system, advanced Large Language
Models (LLMs) have emerged as effective Query Likelihood Models (QLMs) in an
unsupervised way, which re-rank documents based on the probability of
generating the query given the content of a document. However, directly
prompting LLMs to approxi... | [
0.0222334601,
-0.5563320518,
0.77519238,
-0.1135781631,
-0.258322835,
0.0847801715,
0.1696380973,
0.0691133961,
0.1020941436,
0.6132566929,
-0.2570564449,
-0.7869825959,
-0.5431583524,
-0.1200655475,
-0.1784563363,
1.0065574646,
0.2876755297,
-0.094884567,
-0.0968826711,
-0.039... |
http://arxiv.org/abs/2406.13282v1 | http://arxiv.org/pdf/2406.13282v1 | Understanding the RoPE Extensions of Long-Context LLMs: An Attention Perspective | 2024-06-19T07:23:33 | 2024-06-19T07:23:33 | Meizhi Zhong, Chen Zhang, Yikun Lei, Xikai Liu, Yan Gao, Yao Hu, Kehai Chen, Min Zhang | cs.CL | cs.CL | Enabling LLMs to handle lengthy context is currently a research hotspot. Most
LLMs are built upon rotary position embedding (RoPE), a popular position
encoding method. Therefore, a prominent path is to extrapolate the RoPE trained
on comparably short texts to far longer texts. A heavy bunch of efforts have
been dedicat... | [
-0.4976311326,
-0.6048641205,
0.536796689,
0.1317048222,
-0.5588381886,
0.0725781992,
-0.3638248444,
-0.8329967856,
0.3807885945,
0.4050562978,
-0.4751390219,
-0.296082288,
-0.7298557758,
0.0105239935,
-0.5958414078,
1.2456415892,
-0.1911731511,
-0.4922543168,
-0.2243159711,
-0... |
http://arxiv.org/abs/2406.13274v1 | http://arxiv.org/pdf/2406.13274v1 | In-Context Learning on a Budget: A Case Study in Named Entity Recognition | 2024-06-19T07:09:46 | 2024-06-19T07:09:46 | Uri Berger, Tal Baumel, Gabriel Stanovsky | cs.CL | cs.CL | Few shot in-context learning (ICL) typically assumes access to large
annotated training sets. However, in many real world scenarios, such as domain
adaptation, there is only a limited budget to annotate a small number of
samples, with the goal of maximizing downstream performance. We study various
methods for selecting... | [
-0.7876274586,
-0.1357208192,
0.3081864417,
-0.0796065032,
-0.3458629251,
-0.1259656996,
-0.2075574547,
-0.8602828383,
-0.0533362105,
0.3848168254,
-0.501149416,
-0.0853593796,
-0.3518328369,
0.3277324736,
-0.200968951,
1.1674342155,
-0.0286987908,
0.1552905142,
-0.2356358021,
... |
http://arxiv.org/abs/2406.13275v2 | http://arxiv.org/pdf/2406.13275v2 | Enhancing Automated Audio Captioning via Large Language Models with Optimized Audio Encoding | 2024-06-19T07:09:46 | 2024-06-25T08:07:36 | Jizhong Liu, Gang Li, Junbo Zhang, Heinrich Dinkel, Yongqing Wang, Zhiyong Yan, Yujun Wang, Bin Wang | cs.SD | cs.SD, cs.CL, eess.AS | Automated audio captioning (AAC) is an audio-to-text task to describe audio
contents in natural language. Recently, the advancements in large language
models (LLMs), with improvements in training approaches for audio encoders,
have opened up possibilities for improving AAC. Thus, we explore enhancing AAC
from three asp... | [
-0.4472236633,
-0.4639536142,
0.5491535664,
0.0586657077,
-0.5747398138,
0.4450887442,
-0.1921579093,
-0.5523741841,
0.2263264805,
0.3109340668,
-0.2215197086,
-0.6239872575,
-0.8499640822,
0.5131853819,
-0.3968170583,
0.7934072018,
-0.0238248613,
0.2476734221,
-0.0842612088,
-... |
http://arxiv.org/abs/2406.13269v1 | http://arxiv.org/pdf/2406.13269v1 | Investigating Low-Cost LLM Annotation for~Spoken Dialogue Understanding Datasets | 2024-06-19T06:59:57 | 2024-06-19T06:59:57 | Lucas Druart, Valentin Vielzeuf, Yannick Estève | cs.AI | cs.AI, cs.CL, cs.HC, eess.SP | In spoken Task-Oriented Dialogue (TOD) systems, the choice of the semantic
representation describing the users' requests is key to a smooth interaction.
Indeed, the system uses this representation to reason over a database and its
domain knowledge to choose its next action. The dialogue course thus depends on
the infor... | [
-0.5234664083,
-0.6129392982,
0.6266495585,
-0.0950148553,
-0.2297635525,
-0.1341430098,
-0.2147417963,
-0.6312983632,
0.0045911763,
1.0264697075,
-0.8411467671,
-0.5305691957,
-0.3812724352,
0.2011485398,
-0.2766872048,
0.9263228178,
0.2042388916,
0.0468546301,
-0.1037785485,
... |
http://arxiv.org/abs/2406.13261v2 | http://arxiv.org/pdf/2406.13261v2 | BeHonest: Benchmarking Honesty of Large Language Models | 2024-06-19T06:46:59 | 2024-07-01T15:18:07 | Steffi Chern, Zhulin Hu, Yuqing Yang, Ethan Chern, Yuan Guo, Jiahe Jin, Binjie Wang, Pengfei Liu | cs.CL | cs.CL, cs.AI | Previous works on Large Language Models (LLMs) have mainly focused on
evaluating their helpfulness or harmlessness. However, honesty, another crucial
alignment criterion, has received relatively less attention. Dishonest
behaviors in LLMs, such as spreading misinformation and defrauding users,
eroding user trust, and c... | [
0.1307324916,
-0.5545740724,
0.5620586276,
0.3467524052,
-0.1223295107,
0.1633704454,
-0.4884386063,
-0.5650759339,
-0.2859150171,
0.4412038326,
-0.5010533929,
-0.5729340911,
-0.7178253531,
-0.1430512518,
-0.4470469356,
0.8613021374,
-0.0025016356,
0.1738430262,
-0.2135926336,
... |
http://arxiv.org/abs/2406.13250v1 | http://arxiv.org/pdf/2406.13250v1 | LangTopo: Aligning Language Descriptions of Graphs with Tokenized Topological Modeling | 2024-06-19T06:20:22 | 2024-06-19T06:20:22 | Zhong Guan, Hongke Zhao, Likang Wu, Ming He, Jianpin Fan | cs.AI | cs.AI, cs.CL, cs.LG | Recently, large language models (LLMs) have been widely researched in the
field of graph machine learning due to their outstanding abilities in language
comprehension and learning. However, the significant gap between natural
language tasks and topological structure modeling poses a nonnegligible
challenge. Specificall... | [
-0.1780125201,
-0.8603224754,
0.3867532611,
-0.3483709693,
-0.2719884813,
0.3285544217,
-0.4398503006,
-0.3961233795,
0.4144662917,
0.7384839058,
-0.2044991404,
-0.844340086,
-0.6503072977,
0.1026749536,
-0.3157559931,
1.0075467825,
-0.1644732356,
-0.0500469282,
-0.3623993397,
... |
http://arxiv.org/abs/2406.13249v1 | http://arxiv.org/pdf/2406.13249v1 | R^2AG: Incorporating Retrieval Information into Retrieval Augmented Generation | 2024-06-19T06:19:48 | 2024-06-19T06:19:48 | Fuda Ye, Shuangyin Li, Yongqi Zhang, Lei Chen | cs.CL | cs.CL, cs.AI, cs.IR | Retrieval augmented generation (RAG) has been applied in many scenarios to
augment large language models (LLMs) with external documents provided by
retrievers. However, a semantic gap exists between LLMs and retrievers due to
differences in their training objectives and architectures. This misalignment
forces LLMs to p... | [
-0.4275286198,
-0.6943730712,
0.4293416142,
-0.0304700099,
-0.1878511459,
0.2551786602,
-0.1035903767,
-0.6732080579,
0.2856477201,
0.3114230931,
-0.3337857425,
0.1334563196,
-0.5993737578,
0.0688617229,
-0.1784152985,
0.6746014953,
0.0936564952,
-0.0982558131,
-0.3300859034,
-... |
http://arxiv.org/abs/2406.13246v1 | http://arxiv.org/pdf/2406.13246v1 | GSR-BENCH: A Benchmark for Grounded Spatial Reasoning Evaluation via Multimodal LLMs | 2024-06-19T06:15:26 | 2024-06-19T06:15:26 | Navid Rajabi, Jana Kosecka | cs.CL | cs.CL, cs.CV, cs.LG | The ability to understand and reason about spatial relationships between
objects in images is an important component of visual reasoning. This skill
rests on the ability to recognize and localize objects of interest and
determine their spatial relation. Early vision and language models (VLMs) have
been shown to struggl... | [
-0.646732688,
-0.5279172659,
1.0343595743,
0.0508407801,
-0.1421906501,
-0.2766562998,
-0.1355390549,
-0.3427459598,
-0.3698364794,
0.2121573538,
-0.5755441785,
-0.5861105323,
-0.6572245955,
-0.0482692532,
-0.0716576204,
0.4622137547,
-0.0940672085,
0.500362277,
-0.6730778813,
... |
http://arxiv.org/abs/2406.13236v1 | http://arxiv.org/pdf/2406.13236v1 | Data Contamination Can Cross Language Barriers | 2024-06-19T05:53:27 | 2024-06-19T05:53:27 | Feng Yao, Yufan Zhuang, Zihao Sun, Sunan Xu, Animesh Kumar, Jingbo Shang | cs.CL | cs.CL, cs.AI | The opacity in developing large language models (LLMs) is raising growing
concerns about the potential contamination of public benchmarks in the
pre-training data. Existing contamination detection methods are typically based
on the text overlap between training and evaluation data, which can be too
superficial to refle... | [
-0.4301003814,
-0.9012473226,
0.5749421716,
0.0120852375,
-0.3051989377,
0.3585801423,
-0.2356123924,
-0.9952589273,
-0.1845111996,
1.0186767578,
-0.1607231647,
-0.4718319774,
-0.8551653028,
-0.1635692716,
-0.5168709159,
0.8591501117,
-0.3494229019,
-0.1125363782,
-0.1799576581,
... |
http://arxiv.org/abs/2406.13232v1 | http://arxiv.org/pdf/2406.13232v1 | Towards Robust Evaluation: A Comprehensive Taxonomy of Datasets and Metrics for Open Domain Question Answering in the Era of Large Language Models | 2024-06-19T05:43:02 | 2024-06-19T05:43:02 | Akchay Srivastava, Atif Memon | cs.CL | cs.CL, cs.AI, cs.IR, cs.LG | Open Domain Question Answering (ODQA) within natural language processing
involves building systems that answer factual questions using large-scale
knowledge corpora. Recent advances stem from the confluence of several factors,
such as large-scale training datasets, deep learning techniques, and the rise
of large langua... | [
-0.5983136892,
-0.963218689,
0.44837147,
-0.2693030834,
0.1699690223,
0.2951380014,
-0.2771156728,
-0.312148422,
-0.4686228633,
0.2327415645,
-0.4091421366,
-0.3836881518,
-0.3464719653,
0.0328985713,
-0.2279437184,
0.8587871194,
-0.2670994997,
0.1610547602,
-0.5619974136,
-0.5... |
http://arxiv.org/abs/2406.13230v1 | http://arxiv.org/pdf/2406.13230v1 | Enhancing Language Model Factuality via Activation-Based Confidence Calibration and Guided Decoding | 2024-06-19T05:33:34 | 2024-06-19T05:33:34 | Xin Liu, Farima Fatahi Bayat, Lu Wang | cs.CL | cs.CL | Calibrating language models (LMs) aligns their generation confidence with the
actual likelihood of answer correctness, which can inform users about LMs'
reliability and mitigate hallucinated content. However, prior calibration
methods, such as self-consistency-based and logit-based approaches, are either
limited in inf... | [
-0.0883683115,
-0.7217437029,
0.6627195477,
0.057312876,
-0.0500230081,
0.1818494648,
-0.2676279247,
-0.2673324347,
0.0308375806,
0.5431267023,
-0.2100466192,
-0.6046926379,
-0.425632596,
-0.1057433113,
-0.1074680462,
0.7098481059,
0.1349974275,
0.1065683439,
-0.3492696583,
-0.... |
http://arxiv.org/abs/2406.13229v1 | http://arxiv.org/pdf/2406.13229v1 | Probing the Emergence of Cross-lingual Alignment during LLM Training | 2024-06-19T05:31:59 | 2024-06-19T05:31:59 | Hetong Wang, Pasquale Minervini, Edoardo M. Ponti | cs.CL | cs.CL, cs.AI, cs.LG | Multilingual Large Language Models (LLMs) achieve remarkable levels of
zero-shot cross-lingual transfer performance. We speculate that this is
predicated on their ability to align languages without explicit supervision
from parallel sentences. While representations of translationally equivalent
sentences in different l... | [
-0.4366745651,
-0.4817988276,
0.4441693127,
0.3505291939,
0.1054353639,
0.2097680271,
-0.258847177,
-0.5746831298,
0.2618680894,
-0.0834584087,
-0.8301957846,
-0.152323544,
-0.5230927467,
0.0908804089,
-0.4002976716,
0.7575144172,
-0.1767060608,
0.3448372781,
-0.1153224036,
-0.... |
http://arxiv.org/abs/2406.13219v1 | http://arxiv.org/pdf/2406.13219v1 | MC-MKE: A Fine-Grained Multimodal Knowledge Editing Benchmark Emphasizing Modality Consistency | 2024-06-19T05:15:21 | 2024-06-19T05:15:21 | Junzhe Zhang, Huixuan Zhang, Xunjian Yin, Baizhou Huang, Xu Zhang, Xinyu Hu, Xiaojun Wan | cs.CV | cs.CV, cs.CL | Multimodal large language models (MLLMs) are prone to non-factual or outdated
knowledge issues, which can manifest as misreading and misrecognition errors
due to the complexity of multimodal knowledge. Previous benchmarks have not
systematically analyzed the performance of editing methods in correcting these
two error ... | [
-0.3567700088,
-0.7731458545,
0.5750337839,
0.1680753976,
0.1426898837,
0.0873340815,
-0.3664431274,
-0.2633671165,
-0.0281574409,
0.2407883704,
-0.430277586,
-0.4154137671,
-0.6181159019,
0.2875168025,
0.2358995825,
0.5149977207,
-0.3264996111,
0.2660700679,
-0.4956073761,
-0.... |
http://arxiv.org/abs/2406.13217v1 | http://arxiv.org/pdf/2406.13217v1 | Bridging Law and Data: Augmenting Reasoning via a Semi-Structured Dataset with IRAC methodology | 2024-06-19T04:59:09 | 2024-06-19T04:59:09 | Xiaoxi Kang, Lizhen Qu, Lay-Ki Soon, Zhuang Li, Adnan Trakic | cs.CL | cs.CL | The effectiveness of Large Language Models (LLMs) in legal reasoning is often
limited due to the unique legal terminologies and the necessity for highly
specialized knowledge. These limitations highlight the need for high-quality
data tailored for complex legal reasoning tasks. This paper introduces
LEGALSEMI, a benchm... | [
-0.2101056725,
-0.6975775957,
0.6889957786,
0.1046581417,
-0.2578460276,
0.1628296971,
-0.1292335093,
-0.3787648976,
-0.0962440223,
0.7053619623,
-0.2805283368,
-1.0680216551,
-0.2738617957,
0.0057034274,
-0.3864619434,
0.7464946508,
-0.1236140579,
0.123272799,
-0.468419373,
-0... |
http://arxiv.org/abs/2406.13213v1 | http://arxiv.org/pdf/2406.13213v1 | Multi-Meta-RAG: Improving RAG for Multi-Hop Queries using Database Filtering with LLM-Extracted Metadata | 2024-06-19T04:53:48 | 2024-06-19T04:53:48 | Mykhailo Poliakov, Nadiya Shvai | cs.CL | cs.CL, cs.AI, cs.DB | The retrieval-augmented generation (RAG) enables retrieval of relevant
information from an external knowledge source and allows large language models
(LLMs) to answer queries over previously unseen document collections. However,
it was demonstrated that traditional RAG applications perform poorly in
answering multi-hop... | [
-0.8187135458,
-0.7668697834,
0.6107102036,
-0.0502204411,
-0.2496156693,
0.422449708,
0.0585370548,
-0.6624428034,
0.649887383,
0.9268302917,
-0.5890192389,
-0.3566537797,
-0.4901219904,
0.2005129606,
-0.2815147042,
1.0243108273,
0.3113197386,
0.0378348343,
-0.3332919776,
-0.3... |
http://arxiv.org/abs/2406.13195v1 | http://arxiv.org/pdf/2406.13195v1 | Learning Translations via Matrix Completion | 2024-06-19T04:03:19 | 2024-06-19T04:03:19 | Derry Wijaya, Brendan Callahan, John Hewitt, Jie Gao, Xiao Ling, Marianna Apidianaki, Chris Callison-Burch | cs.CL | cs.CL, I.2.7 | Bilingual Lexicon Induction is the task of learning word translations without
bilingual parallel corpora. We model this task as a matrix completion problem,
and present an effective and extendable framework for completing the matrix.
This method harnesses diverse bilingual and monolingual signals, each of which
may be ... | [
-0.4389293492,
-0.1139294058,
0.539329946,
0.2101540565,
-0.2364275157,
0.5554884672,
-0.4116367102,
-0.3313618302,
0.4985563159,
0.4223003685,
-0.8063939214,
-0.5233070254,
-0.7942460179,
0.3869890273,
-0.5155490637,
1.1589305401,
-0.4462237358,
0.3066178858,
-0.3208609223,
-0... |
http://arxiv.org/abs/2406.13193v1 | http://arxiv.org/pdf/2406.13193v1 | PRESTO: Progressive Pretraining Enhances Synthetic Chemistry Outcomes | 2024-06-19T03:59:46 | 2024-06-19T03:59:46 | He Cao, Yanjun Shao, Zhiyuan Liu, Zijing Liu, Xiangru Tang, Yuan Yao, Yu Li | cs.LG | cs.LG, cs.AI, cs.CL, physics.chem-ph | Multimodal Large Language Models (MLLMs) have seen growing adoption across
various scientific disciplines. These advancements encourage the investigation
of molecule-text modeling within synthetic chemistry, a field dedicated to
designing and conducting chemical reactions to synthesize new compounds with
desired proper... | [
-0.3371497989,
-0.8547224402,
0.8973849416,
0.1580500603,
-0.1338986903,
0.5290108323,
-0.313308686,
-0.1911549419,
0.1691804528,
0.3421804011,
-0.7275289893,
-0.6025963426,
-0.4133895636,
0.389252156,
-0.2111769915,
0.9629550576,
-0.378962785,
0.116646193,
-0.430998683,
0.2632... |
http://arxiv.org/abs/2406.13188v1 | http://arxiv.org/pdf/2406.13188v1 | Synthetic Context Generation for Question Generation | 2024-06-19T03:37:52 | 2024-06-19T03:37:52 | Naiming Liu, Zichao Wang, Richard Baraniuk | cs.CL | cs.CL, cs.LG | Despite rapid advancements in large language models (LLMs), QG remains a
challenging problem due to its complicated process, open-ended nature, and the
diverse settings in which question generation occurs. A common approach to
address these challenges involves fine-tuning smaller, custom models using
datasets containin... | [
-0.6023193002,
-1.1255791187,
0.7019549608,
-0.1474743634,
-0.2342503667,
-0.1207640544,
-0.3216225803,
-0.4966432452,
-0.3060725927,
0.4549946785,
-0.8368009925,
-0.1831401885,
-0.1452434063,
0.3980029523,
-0.3239863217,
0.8547973633,
0.0469978862,
0.137717694,
-0.4313186407,
... |
http://arxiv.org/abs/2406.13185v1 | http://arxiv.org/pdf/2406.13185v1 | Learnable In-Context Vector for Visual Question Answering | 2024-06-19T03:33:45 | 2024-06-19T03:33:45 | Yingzhe Peng, Chenduo Hao, Xu Yang, Jiawei Peng, Xinting Hu, Xin Geng | cs.CL | cs.CL | As language models continue to scale, Large Language Models (LLMs) have
exhibited emerging capabilities in In-Context Learning (ICL), enabling them to
solve language tasks by prefixing a few in-context demonstrations (ICDs) as
context. Inspired by these advancements, researchers have extended these
techniques to develo... | [
-0.5149953961,
-0.8327402472,
0.3839928508,
0.1863906682,
-0.4792056382,
0.2459648699,
-0.0947065875,
-0.3900110424,
-0.1797683388,
0.1721577644,
-0.5585845709,
-0.4461408854,
-0.3630961776,
-0.0507509671,
-0.3877971768,
0.6943984032,
-0.1520939469,
0.2123759985,
-0.9510225654,
... |
http://arxiv.org/abs/2406.13184v1 | http://arxiv.org/pdf/2406.13184v1 | Locating and Extracting Relational Concepts in Large Language Models | 2024-06-19T03:29:51 | 2024-06-19T03:29:51 | Zijian Wang, Britney White, Chang Xu | cs.CL | cs.CL | Relational concepts are indeed foundational to the structure of knowledge
representation, as they facilitate the association between various entity
concepts, allowing us to express and comprehend complex world knowledge. By
expressing relational concepts in natural language prompts, people can
effortlessly interact wit... | [
-0.3273397684,
-1.0020363331,
1.1020869017,
0.1286050826,
-0.0583689511,
-0.1116902605,
-0.164758727,
-0.4001210332,
0.3801854551,
0.4150441587,
-0.4400572479,
-0.2890368402,
-0.5793824792,
0.0729026124,
-0.3006101549,
0.5721443295,
-0.151429981,
-0.1969888955,
-0.3978669643,
-... |
http://arxiv.org/abs/2406.13173v2 | http://arxiv.org/pdf/2406.13173v2 | Biomedical Visual Instruction Tuning with Clinician Preference Alignment | 2024-06-19T03:07:33 | 2024-06-30T01:22:09 | Hejie Cui, Lingjun Mao, Xin Liang, Jieyu Zhang, Hui Ren, Quanzheng Li, Xiang Li, Carl Yang | cs.CV | cs.CV, cs.AI, cs.CL, cs.LG, 68T50, 68T45, 68T37, 68T05, 68T07, 68T09,, I.2.7; I.2.6; I.2.10 | Recent advancements in multimodal foundation models have showcased impressive
capabilities in understanding and reasoning with visual and textual
information. Adapting these foundation models trained for general usage to
specialized domains like biomedicine requires large-scale domain-specific
instruction datasets. Whi... | [
-0.1609116644,
-0.7012556195,
0.6656112075,
-0.1553027481,
-0.300222069,
-0.1412600577,
-0.2051730007,
-0.0945084244,
0.0919748992,
0.5237555504,
-0.528401494,
-0.8760365248,
-0.540410161,
-0.3356310725,
0.3922457993,
1.1332455873,
0.0000216986,
0.4021522403,
-0.6808047295,
-0.... |
http://arxiv.org/abs/2406.13170v1 | http://arxiv.org/pdf/2406.13170v1 | Amphista: Accelerate LLM Inference with Bi-directional Multiple Drafting Heads in a Non-autoregressive Style | 2024-06-19T02:53:39 | 2024-06-19T02:53:39 | Zeping Li, Xinlong Yang, Ziheng Gao, Ji Liu, Zhuang Liu, Dong Li, Jinzhang Peng, Lu Tian, Emad Barsoum | cs.AI | cs.AI, cs.CL | Large Language Models (LLMs) inherently use autoregressive decoding, which
lacks parallelism in inference and results in significantly slow inference
speeds, especially when hardware parallel accelerators and memory bandwidth are
not fully utilized. In this work, we propose Amphista, a speculative decoding
algorithm th... | [
-0.2986777425,
-0.877245903,
0.6405568123,
0.5636073947,
-0.5130910873,
0.0516612679,
-0.1261836737,
-0.6656312346,
0.5354393125,
0.5955484509,
-0.7167974114,
-0.3273463845,
-0.6092799306,
0.3057260811,
-0.6155521274,
0.279650867,
-0.0990279838,
-0.1359741241,
0.0301649049,
-0.... |
http://arxiv.org/abs/2406.13167v1 | http://arxiv.org/pdf/2406.13167v1 | QRMeM: Unleash the Length Limitation through Question then Reflection Memory Mechanism | 2024-06-19T02:46:18 | 2024-06-19T02:46:18 | Bo Wang, Heyan Huang, Yixin Cao, Jiahao Ying, Wei Tang, Chong Feng | cs.CL | cs.CL | While large language models (LLMs) have made notable advancements in natural
language processing, they continue to struggle with processing extensive text.
Memory mechanism offers a flexible solution for managing long contexts,
utilizing techniques such as compression, summarization, and structuring to
facilitate nuanc... | [
-0.4681452811,
-0.9817902446,
0.9137724638,
0.1164663211,
-0.6102540493,
0.2277051806,
-0.2103895396,
-0.2755127549,
0.0051932964,
0.6457838416,
-0.5026786923,
-0.3256453872,
-0.2748392522,
0.0351423845,
-0.2007620484,
0.9205696583,
-0.2604107857,
-0.1173386425,
-0.6067008376,
... |
http://arxiv.org/abs/2406.13163v1 | http://arxiv.org/pdf/2406.13163v1 | LLMatDesign: Autonomous Materials Discovery with Large Language Models | 2024-06-19T02:35:02 | 2024-06-19T02:35:02 | Shuyi Jia, Chao Zhang, Victor Fung | cond-mat.mtrl-sci | cond-mat.mtrl-sci, cs.AI, cs.CL | Discovering new materials can have significant scientific and technological
implications but remains a challenging problem today due to the enormity of the
chemical space. Recent advances in machine learning have enabled data-driven
methods to rapidly screen or generate promising materials, but these methods
still depe... | [
-0.0732015595,
-0.5163856745,
1.1820201874,
0.143126592,
0.016567599,
0.4490367174,
-0.1131974161,
-0.2380364835,
0.3177720308,
0.278727442,
-0.8178800344,
-0.6412581205,
-0.281144321,
0.3196629882,
-0.2093406171,
0.4082119465,
-0.5106779933,
0.0490913577,
-0.3054581285,
0.1278... |
http://arxiv.org/abs/2406.13161v1 | http://arxiv.org/pdf/2406.13161v1 | APPL: A Prompt Programming Language for Harmonious Integration of Programs and Large Language Model Prompts | 2024-06-19T02:29:59 | 2024-06-19T02:29:59 | Honghua Dong, Qidong Su, Yubo Gao, Zhaoyu Li, Yangjun Ruan, Gennady Pekhimenko, Chris J. Maddison, Xujie Si | cs.AI | cs.AI, cs.CL, cs.LG, cs.PL | Large Language Models (LLMs) have become increasingly capable of handling
diverse tasks with the aid of well-crafted prompts and integration of external
tools, but as task complexity rises, the workflow involving LLMs can be
complicated and thus challenging to implement and maintain. To address this
challenge, we propo... | [
-0.2589963377,
-0.8121128678,
0.7173924446,
0.4748726785,
0.138201341,
0.4535152018,
-0.2659577727,
-0.1794410199,
-0.1469253451,
0.269807905,
-0.4398142695,
-0.4409996569,
-0.0855532885,
0.289293766,
-0.2738811672,
0.6816808581,
-0.2809167802,
-0.0899085328,
-0.2606997192,
-0.... |
http://arxiv.org/abs/2406.13152v1 | http://arxiv.org/pdf/2406.13152v1 | Analyzing Diversity in Healthcare LLM Research: A Scientometric Perspective | 2024-06-19T02:00:51 | 2024-06-19T02:00:51 | David Restrepo, Chenwei Wu, Constanza Vásquez-Venegas, João Matos, Jack Gallifant, Luis Filipe | cs.CL | cs.CL | The deployment of large language models (LLMs) in healthcare has demonstrated
substantial potential for enhancing clinical decision-making, administrative
efficiency, and patient outcomes. However, the underrepresentation of diverse
groups in the development and application of these models can perpetuate
biases, leadin... | [
0.0544251464,
-0.2050069273,
0.5639145374,
0.3579420149,
0.1068819687,
0.411025703,
-0.1017634943,
-0.7859651446,
0.4259164631,
0.097699374,
-0.4533018768,
-0.4619775712,
-0.8215379119,
0.2619272768,
-0.2296064943,
0.8895866871,
0.2394968569,
0.7449121475,
-0.4237110913,
-0.181... |
http://arxiv.org/abs/2406.13144v1 | http://arxiv.org/pdf/2406.13144v1 | DialSim: A Real-Time Simulator for Evaluating Long-Term Dialogue Understanding of Conversational Agents | 2024-06-19T01:37:10 | 2024-06-19T01:37:10 | Jiho Kim, Woosog Chay, Hyeonji Hwang, Daeun Kyung, Hyunseung Chung, Eunbyeol Cho, Yohan Jo, Edward Choi | cs.CL | cs.CL, cs.AI | Recent advancements in Large Language Models (LLMs) have significantly
enhanced the capabilities of conversational agents, making them applicable to
various fields (e.g., education). Despite their progress, the evaluation of the
agents often overlooks the complexities of real-world conversations, such as
real-time inte... | [
-0.4214900434,
-0.9850693941,
0.4788479507,
0.4181965888,
-0.190078944,
0.4962474406,
-0.2912555039,
-0.3861027956,
0.0980907604,
0.7068821788,
-0.8128674626,
-0.074579373,
-0.2492538393,
0.0115962299,
-0.0911237076,
0.681429565,
0.1809664965,
0.2695392072,
-0.6032789946,
-0.01... |
http://arxiv.org/abs/2406.13138v1 | http://arxiv.org/pdf/2406.13138v1 | Large Language Models are Biased Because They Are Large Language Models | 2024-06-19T01:08:03 | 2024-06-19T01:08:03 | Philip Resnik | cs.CL | cs.CL, cs.AI | This paper's primary goal is to provoke thoughtful discussion about the
relationship between bias and fundamental properties of large language models.
We do this by seeking to convince the reader that harmful biases are an
inevitable consequence arising from the design of any large language model as
LLMs are currently ... | [
-0.3228486478,
-0.8969804049,
0.677805841,
0.3196083009,
-0.0880879387,
-0.3244434297,
-0.4304005802,
-0.5717631578,
-0.00592389,
0.9753312469,
-0.3302936554,
-0.0815340206,
-0.6798105836,
-0.2170664668,
-0.533827424,
1.0926146507,
0.1560637057,
0.3439198732,
-0.2411016077,
-0.... |
http://arxiv.org/abs/2406.13133v1 | http://arxiv.org/pdf/2406.13133v1 | PathoLM: Identifying pathogenicity from the DNA sequence through the Genome Foundation Model | 2024-06-19T00:53:48 | 2024-06-19T00:53:48 | Sajib Acharjee Dip, Uddip Acharjee Shuvo, Tran Chau, Haoqiu Song, Petra Choi, Xuan Wang, Liqing Zhang | cs.CL | cs.CL, cs.LG, q-bio.GN | Pathogen identification is pivotal in diagnosing, treating, and preventing
diseases, crucial for controlling infections and safeguarding public health.
Traditional alignment-based methods, though widely used, are computationally
intense and reliant on extensive reference databases, often failing to detect
novel pathoge... | [
-0.1129891053,
-1.0177028179,
0.430532217,
-0.2413973063,
-0.4460289478,
0.1929396838,
0.0197977927,
-0.2926064134,
0.2350900024,
0.5377191901,
-0.4116646051,
-0.5825084448,
-0.9184706807,
0.2031926215,
-0.2989578247,
0.9079711437,
0.3553864658,
-0.0450843014,
-0.2928966284,
-0... |
http://arxiv.org/abs/2406.13131v2 | http://arxiv.org/pdf/2406.13131v2 | When Parts are Greater Than Sums: Individual LLM Components Can Outperform Full Models | 2024-06-19T00:48:44 | 2024-06-24T15:13:13 | Ting-Yun Chang, Jesse Thomason, Robin Jia | cs.CL | cs.CL | This paper studies in-context learning (ICL) by decomposing the output of
large language models into the individual contributions of attention heads and
MLPs (components). We observe curious components: good-performing ones that
individually do well on a classification task, even when the model performs
poorly; bad-per... | [
-0.1814564615,
-0.3853815794,
0.5331460834,
0.1774830669,
-0.4062541425,
0.087821275,
-0.3399288654,
-0.3566958308,
-0.1446103603,
0.3429383934,
-0.4735248089,
-0.342455864,
-0.6151165366,
-0.0134808002,
-0.460347265,
1.1598594189,
0.0520902425,
0.133323878,
-0.5859152079,
-0.4... |
http://arxiv.org/abs/2406.13124v1 | http://arxiv.org/pdf/2406.13124v1 | Learning to Generate Answers with Citations via Factual Consistency Models | 2024-06-19T00:40:19 | 2024-06-19T00:40:19 | Rami Aly, Zhiqiang Tang, Samson Tan, George Karypis | cs.CL | cs.CL | Large Language Models (LLMs) frequently hallucinate, impeding their
reliability in mission-critical situations. One approach to address this issue
is to provide citations to relevant sources alongside generated content,
enhancing the verifiability of generations. However, citing passages accurately
in answers remains a... | [
-0.3150745928,
-0.8766093254,
1.0542107821,
-0.0659198686,
-0.0464711189,
-0.012921785,
-0.5778851509,
-0.3432776332,
-0.168643713,
0.4907342494,
-0.5537336469,
-0.5398399234,
-0.5921881795,
0.3350843489,
-0.1237515658,
1.1009796858,
0.1655029953,
-0.140013814,
-0.5688394904,
-... |
http://arxiv.org/abs/2406.13121v1 | http://arxiv.org/pdf/2406.13121v1 | Can Long-Context Language Models Subsume Retrieval, RAG, SQL, and More? | 2024-06-19T00:28:58 | 2024-06-19T00:28:58 | Jinhyuk Lee, Anthony Chen, Zhuyun Dai, Dheeru Dua, Devendra Singh Sachan, Michael Boratko, Yi Luan, Sébastien M. R. Arnold, Vincent Perot, Siddharth Dalmia, Hexiang Hu, Xudong Lin, Panupong Pasupat, Aida Amini, Jeremy R. Cole, Sebastian Riedel, Iftekhar Naim, Ming-Wei Chang, Kelvin Guu | cs.CL | cs.CL, cs.AI, cs.IR | Long-context language models (LCLMs) have the potential to revolutionize our
approach to tasks traditionally reliant on external tools like retrieval
systems or databases. Leveraging LCLMs' ability to natively ingest and process
entire corpora of information offers numerous advantages. It enhances
user-friendliness by ... | [
-0.3030155897,
-0.7842031121,
0.4515253305,
0.4132763743,
-0.0515771247,
0.1786974519,
-0.4418520033,
-0.7133530974,
-0.3486898243,
0.4713553488,
-0.4874424636,
-0.3469132781,
-0.1685352474,
0.1866261661,
-0.7416415811,
0.8877741694,
0.3563619852,
-0.1199108362,
-0.4142382443,
... |
http://arxiv.org/abs/2406.13114v1 | http://arxiv.org/pdf/2406.13114v1 | Multi-Stage Balanced Distillation: Addressing Long-Tail Challenges in Sequence-Level Knowledge Distillation | 2024-06-19T00:01:14 | 2024-06-19T00:01:14 | Yuhang Zhou, Jing Zhu, Paiheng Xu, Xiaoyu Liu, Xiyao Wang, Danai Koutra, Wei Ai, Furong Huang | cs.CL | cs.CL, cs.AI | Large language models (LLMs) have significantly advanced various natural
language processing tasks, but deploying them remains computationally
expensive. Knowledge distillation (KD) is a promising solution, enabling the
transfer of capabilities from larger teacher LLMs to more compact student
models. Particularly, sequ... | [
-0.3490498364,
-0.6569213867,
0.7384345531,
0.5770304203,
-0.2496789396,
0.5660043359,
-0.3244397044,
-0.1137542054,
-0.2303368598,
0.4380708039,
-0.5808683634,
-0.3907411993,
-0.75014925,
0.1184782386,
-0.4988828003,
1.0931463242,
-0.0221226662,
-0.088289924,
-0.2252180725,
-0... |
http://arxiv.org/abs/2406.13106v3 | http://arxiv.org/pdf/2406.13106v3 | Accelerating Complex Disease Treatment through Network Medicine and GenAI: A Case Study on Drug Repurposing for Breast Cancer | 2024-06-18T23:40:00 | 2024-06-27T08:28:51 | Ahmed Abdeen Hamed, Tamer E. Fandy | cs.AI | cs.AI, cs.CL, cs.IR, I.2; I.2.6 | The objective of this research is to introduce a network specialized in
predicting drugs that can be repurposed by investigating real-world evidence
sources, such as clinical trials and biomedical literature. Specifically, it
aims to generate drug combination therapies for complex diseases (e.g., cancer,
Alzheimer's). ... | [
-0.1911755204,
-0.8296259046,
0.522331655,
0.0828404278,
0.0606818274,
0.412148118,
0.4121456742,
-0.3892377913,
0.6696144342,
0.9922252297,
-0.7430918813,
-0.3457072675,
-0.7982765436,
-0.1325509697,
-0.1975754499,
0.7605031133,
0.0036893538,
0.4820230305,
0.1598815322,
-0.217... |
http://arxiv.org/abs/2406.13094v1 | http://arxiv.org/pdf/2406.13094v1 | Exploring and Benchmarking the Planning Capabilities of Large Language Models | 2024-06-18T22:57:06 | 2024-06-18T22:57:06 | Bernd Bohnet, Azade Nova, Aaron T Parisi, Kevin Swersky, Katayoon Goshvadi, Hanjun Dai, Dale Schuurmans, Noah Fiedel, Hanie Sedghi | cs.CL | cs.CL, cs.AI, cs.LG | We seek to elevate the planning capabilities of Large Language Models
(LLMs)investigating four main directions. First, we construct a comprehensive
benchmark suite encompassing both classical planning domains and natural
language scenarios. This suite includes algorithms to generate instances with
varying levels of dif... | [
-0.5581745505,
-0.6133156419,
0.7504740357,
0.22697106,
0.0669133738,
-0.0249978025,
-0.5928328633,
-0.5223317742,
-0.2751504183,
0.5500198007,
-0.7004289031,
-0.4927599728,
-0.2375276983,
0.1264376044,
-0.2362384349,
0.3719060123,
-0.2840551138,
0.0863392428,
-0.1019885615,
-0... |
http://arxiv.org/abs/2406.13092v1 | http://arxiv.org/pdf/2406.13092v1 | Multilingual Synopses of Movie Narratives: A Dataset for Story Understanding | 2024-06-18T22:44:50 | 2024-06-18T22:44:50 | Yidan Sun, Jianfei Yu, Boyang Li | cs.CL | cs.CL | Story video-text alignment, a core task in computational story understanding,
aims to align video clips with corresponding sentences in their descriptions.
However, progress on the task has been held back by the scarcity of manually
annotated video-text correspondence and the heavy concentration on English
narrations o... | [
-0.2898439765,
-0.3657028079,
0.3239991069,
0.014101496,
-0.2588741183,
0.3003783226,
-0.3616134226,
-0.3968448341,
0.5186254382,
0.1921296418,
-0.8325922489,
-0.3853601515,
-0.5433793664,
0.4097563624,
-0.166996181,
1.3282754421,
-0.4317684472,
0.2227527201,
-0.0749510974,
-0.... |
http://arxiv.org/abs/2406.13069v2 | http://arxiv.org/pdf/2406.13069v2 | Evaluating $n$-Gram Novelty of Language Models Using Rusty-DAWG | 2024-06-18T21:31:19 | 2024-06-25T15:02:40 | William Merrill, Noah A. Smith, Yanai Elazar | cs.CL | cs.CL, cs.AI | How novel are texts generated by language models (LMs) relative to their
training corpora? In this work, we investigate the extent to which modern LMs
generate $n$-grams from their training data, evaluating both (i) the
probability LMs assign to complete training $n$-grams and (ii) $n$-novelty, the
proportion of $n$-gr... | [
-0.2778188586,
-0.9599656463,
0.7400894761,
-0.2396139801,
-0.3829073906,
0.1542434245,
-0.2253535241,
-0.2811064124,
0.1993981451,
0.1526605189,
-0.4166451991,
-0.6868810058,
-0.4853468537,
0.4103597105,
-0.7763810158,
1.0755039454,
0.3037993014,
-0.0831603855,
-0.3080231845,
... |
http://arxiv.org/abs/2406.13050v1 | http://arxiv.org/pdf/2406.13050v1 | Think-then-Act: A Dual-Angle Evaluated Retrieval-Augmented Generation | 2024-06-18T20:51:34 | 2024-06-18T20:51:34 | Yige Shen, Hao Jiang, Hua Qu, Jihong Zhao | cs.CL | cs.CL | Despite their impressive capabilities, large language models (LLMs) often
face challenges such as temporal misalignment and generating hallucinatory
content. Enhancing LLMs with retrieval mechanisms to fetch relevant information
from external sources offers a promising solution. Inspired by the proverb
"Think twice bef... | [
-0.2753271759,
-0.8059589267,
0.5276864171,
0.0932748765,
-0.2174767703,
0.2165714055,
-0.0682156011,
-0.5232151151,
0.248075515,
0.585916996,
-0.6496876478,
0.2198494077,
-0.211626038,
-0.0564613789,
-0.3982127309,
0.5694482923,
-0.0182751995,
0.3894173503,
-0.051276844,
0.034... |
http://arxiv.org/abs/2406.13035v2 | http://arxiv.org/pdf/2406.13035v2 | D2O: Dynamic Discriminative Operations for Efficient Generative Inference of Large Language Models | 2024-06-18T20:01:51 | 2024-06-23T08:27:48 | Zhongwei Wan, Xinjian Wu, Yu Zhang, Yi Xin, Chaofan Tao, Zhihong Zhu, Xin Wang, Siqi Luo, Jing Xiong, Mi Zhang | cs.CL | cs.CL | Efficient inference in Large Language Models (LLMs) is impeded by the growing
memory demands of key-value (KV) caching, especially for longer sequences.
Traditional KV cache eviction strategies, which prioritize less critical
KV-pairs based on attention scores, often degrade generation quality, leading
to issues such a... | [
-0.0615679249,
-0.7676959634,
0.6128836274,
0.2234187871,
-0.3617526591,
0.2747589648,
-0.1708547026,
-0.2762142718,
-0.2362031639,
0.4279533029,
-0.4158338904,
-0.4713312387,
-0.4559259117,
-0.3193099797,
-0.3333663046,
0.4387403727,
0.0261016879,
-0.1835139394,
-0.3403371572,
... |
http://arxiv.org/abs/2406.13009v1 | http://arxiv.org/pdf/2406.13009v1 | Detecting Errors through Ensembling Prompts (DEEP): An End-to-End LLM Framework for Detecting Factual Errors | 2024-06-18T18:59:37 | 2024-06-18T18:59:37 | Alex Chandler, Devesh Surve, Hui Su | cs.CL | cs.CL, cs.AI | Accurate text summarization is one of the most common and important tasks
performed by Large Language Models, where the costs of human review for an
entire document may be high, but the costs of errors in summarization may be
even greater. We propose Detecting Errors through Ensembling Prompts (DEEP) -
an end-to-end la... | [
0.0589411482,
-0.8615122437,
0.7984182239,
0.3141326904,
-0.0722908378,
0.2073643506,
-0.0665863678,
-0.1855858266,
0.0351319872,
0.5754306316,
-0.6164998412,
-0.5150480866,
-0.5945241451,
0.4631489813,
-0.5109578967,
1.0180208683,
-0.1832405627,
-0.0778888166,
-0.6544541717,
-... |
http://arxiv.org/abs/2406.12998v1 | http://arxiv.org/pdf/2406.12998v1 | Articulatory Encodec: Vocal Tract Kinematics as a Codec for Speech | 2024-06-18T18:38:17 | 2024-06-18T18:38:17 | Cheol Jun Cho, Peter Wu, Tejas S. Prabhune, Dhruv Agarwal, Gopala K. Anumanchipalli | eess.AS | eess.AS, cs.AI, cs.CL, cs.SD | Vocal tract articulation is a natural, grounded control space of speech
production. The spatiotemporal coordination of articulators combined with the
vocal source shapes intelligible speech sounds to enable effective spoken
communication. Based on this physiological grounding of speech, we propose a
new framework of ne... | [
-0.315702796,
-0.6624045372,
0.310582608,
0.0377868824,
-0.125592038,
0.0650405511,
-0.0228175353,
-0.1120529175,
0.8457300663,
0.4588387311,
-0.5358470678,
-0.6365063787,
-0.3238357306,
-0.2761698365,
-0.5356056094,
0.7607138157,
0.2609642744,
0.2009714842,
0.0656719357,
-0.14... |
http://arxiv.org/abs/2406.12997v1 | http://arxiv.org/pdf/2406.12997v1 | Suitability of CCA for Generating Latent State/ Variables in Multi-View Textual Data | 2024-06-18T18:37:24 | 2024-06-18T18:37:24 | Akanksha Mehndiratta, Krishna Asawa | cs.CL | cs.CL | The probabilistic interpretation of Canonical Correlation Analysis (CCA) for
learning low-dimensional real vectors, called as latent variables, has been
exploited immensely in various fields. This study takes a step further by
demonstrating the potential of CCA in discovering a latent state that captures
the contextual... | [
-0.0050493032,
-1.0203288794,
0.9554581046,
0.1462588608,
-0.4151577353,
0.0031084979,
0.1342080235,
-0.1964479834,
0.5345875621,
0.2258487791,
-0.3177340627,
-0.6180227399,
-0.4092802107,
-0.190821901,
-0.2918176055,
0.9832625389,
-0.0505616702,
0.2122288048,
-0.5346719027,
-0... |
http://arxiv.org/abs/2406.12975v1 | http://arxiv.org/pdf/2406.12975v1 | SHIELD: Evaluation and Defense Strategies for Copyright Compliance in LLM Text Generation | 2024-06-18T18:00:03 | 2024-06-18T18:00:03 | Xiaoze Liu, Ting Sun, Tianyang Xu, Feijie Wu, Cunxiang Wang, Xiaoqian Wang, Jing Gao | cs.CL | cs.CL, cs.AI, cs.CY | Large Language Models (LLMs) have transformed machine learning but raised
significant legal concerns due to their potential to produce text that
infringes on copyrights, resulting in several high-profile lawsuits. The legal
landscape is struggling to keep pace with these rapid advancements, with
ongoing debates about w... | [
-0.3518446088,
-0.8294555545,
0.3286593556,
0.1510789245,
-0.2671835423,
0.4274407625,
-0.0805088952,
-0.4867414832,
-0.0861678943,
1.1279051304,
-0.5304871798,
-0.4029420614,
-0.7667841315,
0.2430064976,
-0.6233785748,
1.1506569386,
-0.1763970852,
-0.0444084816,
-0.3189309835,
... |
http://arxiv.org/abs/2406.12845v1 | http://arxiv.org/pdf/2406.12845v1 | Interpretable Preferences via Multi-Objective Reward Modeling and Mixture-of-Experts | 2024-06-18T17:58:28 | 2024-06-18T17:58:28 | Haoxiang Wang, Wei Xiong, Tengyang Xie, Han Zhao, Tong Zhang | cs.LG | cs.LG, cs.CL | Reinforcement learning from human feedback (RLHF) has emerged as the primary
method for aligning large language models (LLMs) with human preferences. The
RLHF process typically starts by training a reward model (RM) using human
preference data. Conventional RMs are trained on pairwise responses to the same
user request... | [
-0.4742850065,
-0.3387283385,
0.4888270795,
0.1263673156,
-0.2304513603,
0.1233774424,
-0.425347954,
-0.593172431,
0.1492993832,
0.6052300334,
-0.6091064215,
-0.3080288172,
-0.6947807074,
-0.0395773053,
0.1392645091,
1.1342562437,
-0.0587643832,
-0.0941063389,
-0.0965203792,
-0... |
http://arxiv.org/abs/2406.12832v1 | http://arxiv.org/pdf/2406.12832v1 | LaMDA: Large Model Fine-Tuning via Spectrally Decomposed Low-Dimensional Adaptation | 2024-06-18T17:52:59 | 2024-06-18T17:52:59 | Seyedarmin Azizi, Souvik Kundu, Massoud Pedram | cs.CL | cs.CL, cs.AI, cs.LG | Low-rank adaptation (LoRA) has become the default approach to fine-tune large
language models (LLMs) due to its significant reduction in trainable
parameters. However, trainable parameter demand for LoRA increases with
increasing model embedding dimensions, leading to high compute costs.
Additionally, its backward upda... | [
-0.6814360619,
-0.8504337668,
0.2876108587,
0.0030407312,
-0.5717936158,
-0.0885734037,
-0.527949512,
-0.38648054,
0.1756724268,
0.6900543571,
-0.3676815927,
-0.5758179426,
-0.4335961342,
-0.4396010339,
-0.0103873732,
0.8046544194,
0.0959328935,
-0.1982876211,
0.058994092,
-0.2... |
http://arxiv.org/abs/2406.12830v1 | http://arxiv.org/pdf/2406.12830v1 | What Are the Odds? Language Models Are Capable of Probabilistic Reasoning | 2024-06-18T17:51:24 | 2024-06-18T17:51:24 | Akshay Paruchuri, Jake Garrison, Shun Liao, John Hernandez, Jacob Sunshine, Tim Althoff, Xin Liu, Daniel McDuff | cs.CL | cs.CL | Language models (LM) are capable of remarkably complex linguistic tasks;
however, numerical reasoning is an area in which they frequently struggle. An
important but rarely evaluated form of reasoning is understanding probability
distributions. In this paper, we focus on evaluating the probabilistic
reasoning capabiliti... | [
-0.3403833508,
-0.7543029785,
1.022244215,
0.2774446011,
-0.2661717832,
-0.3947249055,
-0.2583685815,
-0.1006497368,
-0.3328454196,
0.7042657733,
-0.4396134317,
-0.6084209681,
-0.212174505,
-0.2133410573,
-0.3372918665,
0.6854724884,
0.0759521127,
0.0711470172,
-0.273848176,
0.... |
http://arxiv.org/abs/2406.12824v1 | http://arxiv.org/pdf/2406.12824v1 | From RAGs to rich parameters: Probing how language models utilize external knowledge over parametric information for factual queries | 2024-06-18T17:46:08 | 2024-06-18T17:46:08 | Hitesh Wadhwa, Rahul Seetharaman, Somyaa Aggarwal, Reshmi Ghosh, Samyadeep Basu, Soundararajan Srinivasan, Wenlong Zhao, Shreyas Chaudhari, Ehsan Aghazadeh | cs.CL | cs.CL, cs.AI | Retrieval Augmented Generation (RAG) enriches the ability of language models
to reason using external context to augment responses for a given user prompt.
This approach has risen in popularity due to practical applications in various
applications of language models in search, question/answering, and chat-bots.
However... | [
-0.4662450254,
-0.963978529,
0.7728066444,
-0.1255215257,
-0.1963133961,
-0.2015936375,
-0.060631223,
-0.2520051599,
0.1724565476,
0.7700517774,
-0.4264701605,
0.1323728412,
-0.5846115351,
-0.0267072357,
-0.4087224007,
1.3168268204,
0.1395659894,
-0.0917904973,
-0.2889223993,
-... |
http://arxiv.org/abs/2406.12822v1 | http://arxiv.org/pdf/2406.12822v1 | Is It Good Data for Multilingual Instruction Tuning or Just Bad Multilingual Evaluation for Large Language Models? | 2024-06-18T17:43:47 | 2024-06-18T17:43:47 | Pinzhen Chen, Simon Yu, Zhicheng Guo, Barry Haddow | cs.CL | cs.CL, cs.AI | Large language models, particularly multilingual ones, are designed, claimed,
and expected to cater to native speakers of varied languages. We hypothesise
that the current practices of fine-tuning and evaluating these models may
mismatch this intention owing to a heavy reliance on translation, which can
introduce trans... | [
-0.3027856052,
-0.7194345593,
0.0702209324,
0.2524472773,
-0.2635110617,
-0.0925595835,
-0.7467886209,
-0.5149406195,
-0.2557646334,
0.4096815586,
-0.3536449969,
-0.5687388778,
-0.4303764999,
0.0925701037,
-0.2384689897,
1.2590044737,
-0.1829913259,
0.3329354525,
-0.1628143787,
... |
http://arxiv.org/abs/2406.12814v1 | http://arxiv.org/pdf/2406.12814v1 | Adversarial Attacks on Multimodal Agents | 2024-06-18T17:32:48 | 2024-06-18T17:32:48 | Chen Henry Wu, Jing Yu Koh, Ruslan Salakhutdinov, Daniel Fried, Aditi Raghunathan | cs.LG | cs.LG, cs.CL, cs.CR, cs.CV | Vision-enabled language models (VLMs) are now used to build autonomous
multimodal agents capable of taking actions in real environments. In this
paper, we show that multimodal agents raise new safety risks, even though
attacking agents is more challenging than prior attacks due to limited access
to and knowledge about ... | [
-0.5641281009,
-0.8847674131,
0.1799155474,
-0.0650162324,
-0.1063222662,
0.2168707401,
0.1370333433,
-0.4260867536,
-0.0618007518,
0.7831940055,
-0.4007405937,
-0.0983343422,
-0.7871780396,
-0.058364369,
-0.5540648699,
1.0394991636,
0.1564160734,
0.0898517519,
-0.1177503914,
-... |
http://arxiv.org/abs/2406.12809v1 | http://arxiv.org/pdf/2406.12809v1 | Can Large Language Models Always Solve Easy Problems if They Can Solve Harder Ones? | 2024-06-18T17:25:47 | 2024-06-18T17:25:47 | Zhe Yang, Yichang Zhang, Tianyu Liu, Jian Yang, Junyang Lin, Chang Zhou, Zhifang Sui | cs.CL | cs.CL | Large language models (LLMs) have demonstrated impressive capabilities, but
still suffer from inconsistency issues (e.g. LLMs can react differently to
disturbances like rephrasing or inconsequential order change). In addition to
these inconsistencies, we also observe that LLMs, while capable of solving hard
problems, c... | [
-0.1240049675,
-0.9583745003,
0.8666974306,
0.3753814995,
0.1726868153,
-0.0159041993,
-0.6034799218,
-0.5053851604,
-0.3734549284,
0.2218021452,
-0.381318748,
-0.3119009435,
-0.844587028,
0.2204431444,
-0.2963439226,
0.8685318232,
-0.2501243353,
-0.0580419861,
-0.1809535027,
-... |
http://arxiv.org/abs/2406.12793v1 | http://arxiv.org/pdf/2406.12793v1 | ChatGLM: A Family of Large Language Models from GLM-130B to GLM-4 All Tools | 2024-06-18T16:58:21 | 2024-06-18T16:58:21 | Team GLM, :, Aohan Zeng, Bin Xu, Bowen Wang, Chenhui Zhang, Da Yin, Diego Rojas, Guanyu Feng, Hanlin Zhao, Hanyu Lai, Hao Yu, Hongning Wang, Jiadai Sun, Jiajie Zhang, Jiale Cheng, Jiayi Gui, Jie Tang, Jing Zhang, Juanzi Li, Lei Zhao, Lindong Wu, Lucen Zhong, Mingdao Liu, Minlie Huang, Peng Zhang, Qinkai Zheng, Rui Lu,... | cs.CL | cs.CL | We introduce ChatGLM, an evolving family of large language models that we
have been developing over time. This report primarily focuses on the GLM-4
language series, which includes GLM-4, GLM-4-Air, and GLM-4-9B. They represent
our most capable models that are trained with all the insights and lessons
gained from the p... | [
-0.2519461215,
-0.6652829051,
0.4075682461,
0.2101272345,
0.1656419486,
0.3329312503,
-0.2095591575,
-0.7486205697,
-0.3720740676,
0.470588237,
-0.3880741298,
-0.5024772286,
-0.4682002366,
-0.0855326131,
-0.3731310666,
0.9715225101,
0.0650931522,
0.1729331017,
-0.1573032141,
-0... |
http://arxiv.org/abs/2406.12787v1 | http://arxiv.org/pdf/2406.12787v1 | Generating Educational Materials with Different Levels of Readability using LLMs | 2024-06-18T16:55:10 | 2024-06-18T16:55:10 | Chieh-Yang Huang, Jing Wei, Ting-Hao 'Kenneth' Huang | cs.CL | cs.CL, cs.HC | This study introduces the leveled-text generation task, aiming to rewrite
educational materials to specific readability levels while preserving meaning.
We assess the capability of GPT-3.5, LLaMA-2 70B, and Mixtral 8x7B, to generate
content at various readability levels through zero-shot and few-shot prompting.
Evaluat... | [
-0.242930755,
-0.890824914,
0.5340454578,
-0.0739474446,
-0.385106355,
0.2339546978,
0.0330212452,
-0.4709134698,
-0.1933537275,
0.2031377554,
-0.6151809692,
-0.7405675054,
-0.6929391026,
0.3838273585,
-0.6103785634,
1.2116354704,
-0.3341094553,
-0.3361873031,
-0.0077142292,
0.... |
http://arxiv.org/abs/2406.12784v1 | http://arxiv.org/pdf/2406.12784v1 | UBENCH: Benchmarking Uncertainty in Large Language Models with Multiple Choice Questions | 2024-06-18T16:50:38 | 2024-06-18T16:50:38 | Xunzhi Wang, Zhuowei Zhang, Qiongyu Li, Gaonan Chen, Mengting Hu, Zhiyu li, Bitong Luo, Hang Gao, Zhixin Han, Haotian Wang | cs.CL | cs.CL | The rapid development of large language models (LLMs) has shown promising
practical results. However, their low interpretability often leads to errors in
unforeseen circumstances, limiting their utility. Many works have focused on
creating comprehensive evaluation systems, but previous benchmarks have
primarily assesse... | [
-0.329413414,
-0.85753721,
0.9105125666,
0.3319861889,
-0.1695575714,
0.170815751,
-0.4473421574,
-0.3459208012,
-0.2766637504,
0.3095332086,
-0.1999500245,
-0.1730401814,
-0.3782480359,
-0.0574469827,
-0.3335809708,
0.48853001,
-0.1219385862,
0.3337317705,
-0.380602181,
-0.238... |
http://arxiv.org/abs/2406.12779v1 | http://arxiv.org/pdf/2406.12779v1 | Composited-Nested-Learning with Data Augmentation for Nested Named Entity Recognition | 2024-06-18T16:46:18 | 2024-06-18T16:46:18 | Xingming Liao, Nankai Lin, Haowen Li, Lianglun Cheng, Zhuowei Wang, Chong Chen | cs.CL | cs.CL | Nested Named Entity Recognition (NNER) focuses on addressing overlapped
entity recognition. Compared to Flat Named Entity Recognition (FNER), annotated
resources are scarce in the corpus for NNER. Data augmentation is an effective
approach to address the insufficient annotated corpus. However, there is a
significant la... | [
-0.4770391285,
-0.3545203805,
0.063658841,
0.1394805312,
-0.4187935889,
0.696143806,
-0.3545637131,
-0.7453289032,
0.1377706975,
0.2861091793,
-0.4615258873,
-0.6953191757,
-0.4785585999,
0.4190247357,
-0.0334409773,
1.0875397921,
-0.0181941818,
0.174357757,
-0.5253830552,
0.02... |
http://arxiv.org/abs/2406.12775v1 | http://arxiv.org/pdf/2406.12775v1 | Hopping Too Late: Exploring the Limitations of Large Language Models on Multi-Hop Queries | 2024-06-18T16:44:13 | 2024-06-18T16:44:13 | Eden Biran, Daniela Gottesman, Sohee Yang, Mor Geva, Amir Globerson | cs.CL | cs.CL | Large language models (LLMs) can solve complex multi-step problems, but
little is known about how these computations are implemented internally.
Motivated by this, we study how LLMs answer multi-hop queries such as "The
spouse of the performer of Imagine is". These queries require two information
extraction steps: a la... | [
-0.4505207241,
-0.8529507518,
0.933884263,
0.5257951617,
0.0777427927,
0.4492661357,
-0.0055363905,
-0.632242322,
0.2215741724,
0.7259854674,
-0.9319960475,
-0.0504340306,
-0.6062455177,
-0.2133720368,
-0.1053932235,
0.7006527185,
-0.0550922751,
-0.2218435854,
-0.1356334984,
0.... |
http://arxiv.org/abs/2406.12770v2 | http://arxiv.org/pdf/2406.12770v2 | Informatics & dairy industry coalition: AI trends and present challenges | 2024-06-18T16:39:21 | 2024-06-19T11:49:03 | Silvia García-Méndez, Francisco de Arriba-Pérez, María del Carmen Somoza-López | cs.AI | cs.AI, cs.CL | Artificial Intelligence (AI) can potentially transform the industry,
enhancing the production process and minimizing manual, repetitive tasks.
Accordingly, the synergy between high-performance computing and powerful
mathematical models enables the application of sophisticated data analysis
procedures like Machine Learn... | [
-0.2390540838,
-0.6871712208,
0.2931370139,
-0.1100316867,
0.185153082,
-0.0426885597,
0.2918942571,
-1.1121965647,
-0.0874698609,
0.5125900507,
-1.0424218178,
-0.3027247488,
-0.8397393227,
0.3990011513,
-0.615131557,
0.7463873625,
-0.1251197159,
-0.0145081207,
-0.4394178987,
-... |
http://arxiv.org/abs/2406.12754v1 | http://arxiv.org/pdf/2406.12754v1 | Chumor 1.0: A Truly Funny and Challenging Chinese Humor Understanding Dataset from Ruo Zhi Ba | 2024-06-18T16:22:05 | 2024-06-18T16:22:05 | Ruiqi He, Yushu He, Longju Bai, Jiarui Liu, Zhenjie Sun, Zenghao Tang, He Wang, Hanchen Xia, Naihao Deng | cs.CL | cs.CL, cs.AI | Existing humor datasets and evaluations predominantly focus on English,
lacking resources for culturally nuanced humor in non-English languages like
Chinese. To address this gap, we construct Chumor, a dataset sourced from Ruo
Zhi Ba (RZB), a Chinese Reddit-like platform dedicated to sharing
intellectually challenging ... | [
-0.0027440337,
-0.7110546231,
0.1804534644,
0.5308893323,
-0.7412741184,
-0.225938186,
-0.1694910973,
-0.3617395163,
0.1330912262,
0.3590416014,
-0.8902884722,
-0.0849268734,
-0.1798070967,
0.191635862,
0.0266695563,
1.2469418049,
0.3781839311,
0.3912275434,
-0.2027166337,
-0.0... |
http://arxiv.org/abs/2406.12753v1 | http://arxiv.org/pdf/2406.12753v1 | OlympicArena: Benchmarking Multi-discipline Cognitive Reasoning for Superintelligent AI | 2024-06-18T16:20:53 | 2024-06-18T16:20:53 | Zhen Huang, Zengzhi Wang, Shijie Xia, Xuefeng Li, Haoyang Zou, Ruijie Xu, Run-Ze Fan, Lyumanshan Ye, Ethan Chern, Yixin Ye, Yikai Zhang, Yuqing Yang, Ting Wu, Binjie Wang, Shichao Sun, Yang Xiao, Yiyuan Li, Fan Zhou, Steffi Chern, Yiwei Qin, Yan Ma, Jiadi Su, Yixiu Liu, Yuxiang Zheng, Shaoting Zhang, Dahua Lin, Yu Qiao... | cs.CL | cs.CL, cs.AI | The evolution of Artificial Intelligence (AI) has been significantly
accelerated by advancements in Large Language Models (LLMs) and Large
Multimodal Models (LMMs), gradually showcasing potential cognitive reasoning
abilities in problem-solving and scientific discovery (i.e., AI4Science) once
exclusive to human intelle... | [
-0.5549191236,
-0.5743791461,
0.9837358594,
0.1628739089,
0.0427730381,
0.0580209941,
-0.3195518553,
-0.8100682497,
-0.2795290351,
0.2126732171,
-0.6063630581,
-0.3138281107,
-0.7827489972,
0.3136721551,
-0.2569106817,
0.7806038857,
-0.146941185,
0.3227481842,
-0.6133568287,
-0... |
http://arxiv.org/abs/2406.12746v3 | http://arxiv.org/pdf/2406.12746v3 | Rationale-based Ensemble of Multiple QA Strategies for Zero-shot Knowledge-based VQA | 2024-06-18T16:06:38 | 2024-06-23T03:06:42 | Miaoyu Li, Haoxin Li, Zilin Du, Boyang Li | cs.CL | cs.CL | Knowledge-based Visual Qustion-answering (K-VQA) necessitates the use of
background knowledge beyond what is depicted in the image. Current zero-shot
K-VQA methods usually translate an image to a single type of textual decision
context and use a text-based model to answer the question based on it, which
conflicts with ... | [
-0.5978816152,
-0.9319287539,
0.93240273,
-0.0206321459,
-0.1946876496,
0.2127022743,
0.1039951667,
-0.2002800703,
0.1582599431,
0.408600688,
-0.991646111,
-0.453289777,
-0.0480778478,
0.3097380996,
-0.0716002434,
0.6100580096,
-0.2258901,
0.0749251246,
-0.4799111784,
-0.624007... |
http://arxiv.org/abs/2406.12742v1 | http://arxiv.org/pdf/2406.12742v1 | Benchmarking Multi-Image Understanding in Vision and Language Models: Perception, Knowledge, Reasoning, and Multi-Hop Reasoning | 2024-06-18T16:02:18 | 2024-06-18T16:02:18 | Bingchen Zhao, Yongshuo Zong, Letian Zhang, Timothy Hospedales | cs.CV | cs.CV, cs.AI, cs.CL | The advancement of large language models (LLMs) has significantly broadened
the scope of applications in natural language processing, with multi-modal LLMs
extending these capabilities to integrate and interpret visual data. However,
existing benchmarks for visual language models (VLMs) predominantly focus on
single-im... | [
-0.6931501031,
-0.6569930315,
0.8541707397,
0.1941063255,
-0.3930422068,
-0.2038321942,
-0.045682475,
-0.6592306495,
-0.4141788781,
0.4031482935,
-0.442230165,
-0.5637641549,
-0.5681396127,
0.0928616375,
-0.0729635656,
0.2836843431,
-0.0557120144,
0.4125419855,
-0.6082412601,
-... |
http://arxiv.org/abs/2406.12739v1 | http://arxiv.org/pdf/2406.12739v1 | Self-Distillation for Model Stacking Unlocks Cross-Lingual NLU in 200+ Languages | 2024-06-18T16:00:20 | 2024-06-18T16:00:20 | Fabian David Schmidt, Philipp Borchert, Ivan Vulić, Goran Glavaš | cs.CL | cs.CL | LLMs have become a go-to solution not just for text generation, but also for
natural language understanding (NLU) tasks. Acquiring extensive knowledge
through language modeling on web-scale corpora, they excel on English NLU, yet
struggle to extend their NLU capabilities to underrepresented languages. In
contrast, mach... | [
-0.2749707103,
-0.3712501228,
0.3989258409,
0.3452984989,
-0.308976233,
0.3903912306,
-0.1010776982,
-0.4145878553,
0.00931268,
0.6051955819,
-0.5263268352,
-0.3481300473,
-0.5795573592,
0.5176107883,
-0.2669531107,
1.0850845575,
-0.1458499283,
0.2233676761,
-0.3998829126,
-0.0... |
http://arxiv.org/abs/2406.12738v1 | http://arxiv.org/pdf/2406.12738v1 | Large Language Model as a Universal Clinical Multi-task Decoder | 2024-06-18T15:58:36 | 2024-06-18T15:58:36 | Yujiang Wu, Hongjian Song, Jiawen Zhang, Xumeng Wen, Shun Zheng, Jiang Bian | cs.CL | cs.CL, cs.AI | The development of effective machine learning methodologies for enhancing the
efficiency and accuracy of clinical systems is crucial. Despite significant
research efforts, managing a plethora of diversified clinical tasks and
adapting to emerging new tasks remain significant challenges. This paper
presents a novel para... | [
0.1000367925,
-0.5882712603,
0.6032119989,
0.2794181108,
-0.361489892,
-0.3728189468,
-0.4915854633,
-0.1067036539,
0.0219854899,
0.5667129755,
-0.4046459794,
-0.6308941245,
-0.9989757538,
0.1512538344,
-0.1647501737,
0.9629340768,
-0.0309398901,
0.5020872951,
0.0272083301,
-0.... |
http://arxiv.org/abs/2406.12725v1 | http://arxiv.org/pdf/2406.12725v1 | Can Large Language Models Code Like a Linguist?: A Case Study in Low Resource Sound Law Induction | 2024-06-18T15:46:04 | 2024-06-18T15:46:04 | Atharva Naik, Kexun Zhang, Nathaniel Robinson, Aravind Mysore, Clayton Marr, Hong Sng Rebecca Byrnes, Anna Cai, Kalvin Chang, David Mortensen | cs.CL | cs.CL, cs.AI | Historical linguists have long written a kind of incompletely formalized
''program'' that converts reconstructed words in an ancestor language into
words in one of its attested descendants that consist of a series of ordered
string rewrite functions (called sound laws). They do this by observing pairs
of words in the r... | [
0.1467373967,
-0.5834323764,
0.6769607663,
0.0371783711,
-0.0291518159,
-0.145380035,
-0.4943360686,
-0.4629685283,
0.301864624,
0.6212772131,
-0.6305488348,
-0.1407812238,
-0.322752893,
0.0878929421,
-0.5751240849,
0.708951354,
0.2270610929,
0.0041291728,
0.2436504215,
-0.2342... |
http://arxiv.org/abs/2406.12719v1 | http://arxiv.org/pdf/2406.12719v1 | On the Robustness of Language Models for Tabular Question Answering | 2024-06-18T15:41:15 | 2024-06-18T15:41:15 | Kushal Raj Bhandari, Sixue Xing, Soham Dan, Jianxi Gao | cs.CL | cs.CL, cs.AI | Large Language Models (LLMs), originally shown to ace various text
comprehension tasks have also remarkably been shown to tackle table
comprehension tasks without specific training. While previous research has
explored LLM capabilities with tabular dataset tasks, our study assesses the
influence of $\textit{in-context ... | [
-0.0455213897,
-1.1748235226,
0.6852290034,
0.0446443371,
-0.1628693491,
-0.0146389259,
-0.0959310532,
-0.2835902274,
-0.4341961443,
0.5605060458,
-0.3107777536,
-0.5928237438,
-0.1557411999,
0.0290951636,
-0.5478314161,
1.088085413,
-0.1884958744,
0.1744638234,
-0.5434870124,
... |
http://arxiv.org/abs/2406.12718v2 | http://arxiv.org/pdf/2406.12718v2 | AGLA: Mitigating Object Hallucinations in Large Vision-Language Models with Assembly of Global and Local Attention | 2024-06-18T15:38:41 | 2024-06-21T17:33:21 | Wenbin An, Feng Tian, Sicong Leng, Jiahao Nie, Haonan Lin, QianYing Wang, Guang Dai, Ping Chen, Shijian Lu | cs.CV | cs.CV, cs.AI, cs.CL | Despite their great success across various multimodal tasks, Large
Vision-Language Models (LVLMs) are facing a prevalent problem with object
hallucinations, where the generated textual responses are inconsistent with
ground-truth objects in the given image. This paper investigates various LVLMs
and pinpoints attention ... | [
-0.5161185265,
-0.9556037784,
0.4679366946,
0.0010124793,
-0.1416337341,
-0.1613081545,
0.1245303378,
-0.828217268,
0.3950973153,
0.5880650282,
-0.3002540767,
-0.1972730309,
-0.3611121178,
0.1145113036,
-0.4961113036,
0.5360131264,
-0.0611172542,
-0.0724547431,
-0.4450871348,
-... |
http://arxiv.org/abs/2406.12708v1 | http://arxiv.org/pdf/2406.12708v1 | AgentReview: Exploring Peer Review Dynamics with LLM Agents | 2024-06-18T15:22:12 | 2024-06-18T15:22:12 | Yiqiao Jin, Qinlin Zhao, Yiyang Wang, Hao Chen, Kaijie Zhu, Yijia Xiao, Jindong Wang | cs.CL | cs.CL | Peer review is fundamental to the integrity and advancement of scientific
publication. Traditional methods of peer review analyses often rely on
exploration and statistics of existing peer review data, which do not
adequately address the multivariate nature of the process, account for the
latent variables, and are furt... | [
-0.1690903306,
-0.6716127992,
0.6628857851,
0.2243362814,
-0.0363293886,
0.2702712119,
0.1282831281,
-0.4337542653,
0.3699555993,
0.4397924542,
-0.4377846718,
-0.3696678281,
-0.6384881139,
0.1692275852,
-0.2282402217,
1.0465073586,
-0.0387618653,
0.5291012526,
0.0106288167,
0.1... |
http://arxiv.org/abs/2406.12707v1 | http://arxiv.org/pdf/2406.12707v1 | Talk With Human-like Agents: Empathetic Dialogue Through Perceptible Acoustic Reception and Reaction | 2024-06-18T15:19:51 | 2024-06-18T15:19:51 | Haoqiu Yan, Yongxin Zhu, Kai Zheng, Bing Liu, Haoyu Cao, Deqiang Jiang, Linli Xu | cs.CL | cs.CL, cs.AI, cs.SD, eess.AS | Large Language Model (LLM)-enhanced agents become increasingly prevalent in
Human-AI communication, offering vast potential from entertainment to
professional domains. However, current multi-modal dialogue systems overlook
the acoustic information present in speech, which is crucial for understanding
human communicatio... | [
-0.2761751115,
-0.7769811749,
0.8676675558,
0.3914391696,
-0.1532315165,
0.0669050589,
-0.3239983022,
-0.4307463765,
0.4554218948,
0.2530001402,
-0.7315779328,
-0.3422760963,
-0.2489989698,
0.165763557,
-0.0249271467,
0.4641507566,
0.352627039,
0.1310400963,
0.2434010506,
-0.02... |
http://arxiv.org/abs/2406.12702v2 | http://arxiv.org/pdf/2406.12702v2 | [WIP] Jailbreak Paradox: The Achilles' Heel of LLMs | 2024-06-18T15:14:35 | 2024-06-21T00:05:35 | Abhinav Rao, Monojit Choudhury, Somak Aditya | cs.CL | cs.CL | We introduce two paradoxes concerning jailbreak of foundation models: First,
it is impossible to construct a perfect jailbreak classifier, and second, a
weaker model cannot consistently detect whether a stronger (in a
pareto-dominant sense) model is jailbroken or not. We provide formal proofs for
these paradoxes and a ... | [
-0.1482927501,
-0.5590061545,
0.9786558747,
0.0813154206,
-0.1834219992,
0.528265059,
0.3642432392,
-0.7766568065,
0.1425966173,
0.6650443077,
-0.2488324344,
0.0172927957,
-0.8224560618,
0.1196168661,
-0.7396988273,
1.0492624044,
-0.2722165883,
0.0667989552,
-0.1872653514,
-0.3... |
http://arxiv.org/abs/2406.12692v1 | http://arxiv.org/pdf/2406.12692v1 | MAGIC: Generating Self-Correction Guideline for In-Context Text-to-SQL | 2024-06-18T15:06:06 | 2024-06-18T15:06:06 | Arian Askari, Christian Poelitz, Xinye Tang | cs.CL | cs.CL, cs.AI, cs.DB, cs.HC | Self-correction in text-to-SQL is the process of prompting large language
model (LLM) to revise its previously incorrectly generated SQL, and commonly
relies on manually crafted self-correction guidelines by human experts that are
not only labor-intensive to produce but also limited by the human ability in
identifying ... | [
-0.064440757,
-0.7444496155,
0.348352015,
0.2678665519,
0.0339630507,
-0.1081962585,
-0.2259566337,
-0.0666861683,
0.3064178824,
1.1127383709,
-0.9572529197,
-0.3942534328,
-0.0794663429,
0.5885931253,
-0.1776633412,
1.3910256624,
0.0421380848,
0.1407071203,
-0.1260834783,
0.04... |
http://arxiv.org/abs/2406.12687v1 | http://arxiv.org/pdf/2406.12687v1 | Using LLMs to Aid Annotation and Collection of Clinically-Enriched Data in Bipolar Disorder and Schizophrenia | 2024-06-18T15:00:24 | 2024-06-18T15:00:24 | Ankit Aich, Avery Quynh, Pamela Osseyi, Amy Pinkham, Philip Harvey, Brenda Curtis, Colin Depp, Natalie Parde | cs.CL | cs.CL | NLP in mental health has been primarily social media focused. Real world
practitioners also have high case loads and often domain specific variables, of
which modern LLMs lack context. We take a dataset made by recruiting 644
participants, including individuals diagnosed with Bipolar Disorder (BD),
Schizophrenia (SZ), ... | [
-0.3449556231,
-0.7388136387,
0.9796881676,
0.3033994734,
-0.1853338778,
-0.0430264547,
-0.1858498305,
-0.5750687718,
0.5113955736,
0.5545511246,
-0.6452884674,
-0.7728431225,
-0.7551493645,
0.3073183596,
-0.4145584702,
1.2147784233,
-0.0303391274,
0.4167316854,
-0.0813454092,
... |
http://arxiv.org/abs/2406.12680v1 | http://arxiv.org/pdf/2406.12680v1 | Measuring Psychological Depth in Language Models | 2024-06-18T14:51:54 | 2024-06-18T14:51:54 | Fabrice Harel-Canada, Hanyu Zhou, Sreya Mupalla, Zeynep Yildiz, Amit Sahai, Nanyun Peng | cs.CL | cs.CL | Evaluations of creative stories generated by large language models (LLMs)
often focus on objective properties of the text, such as its style, coherence,
and toxicity. While these metrics are indispensable, they do not speak to a
story's subjective, psychological impact from a reader's perspective. We
introduce the Psyc... | [
-0.5659772754,
-0.5666450858,
0.9016594887,
0.7922425866,
-0.3148549497,
0.2176250219,
-0.2703851163,
-0.6326383948,
0.0908886641,
0.1272573173,
-0.4257556498,
-0.5526247025,
-0.6768557429,
0.2983304858,
-0.2942495942,
0.9404600859,
-0.0982040316,
-0.19214724,
-0.2760140598,
0.... |
http://arxiv.org/abs/2406.12679v1 | http://arxiv.org/pdf/2406.12679v1 | Vernacular? I Barely Know Her: Challenges with Style Control and Stereotyping | 2024-06-18T14:51:30 | 2024-06-18T14:51:30 | Ankit Aich, Tingting Liu, Salvatore Giorgi, Kelsey Isman, Lyle Ungar, Brenda Curtis | cs.CL | cs.CL | Large Language Models (LLMs) are increasingly being used in educational and
learning applications. Research has demonstrated that controlling for style, to
fit the needs of the learner, fosters increased understanding, promotes
inclusion, and helps with knowledge distillation. To understand the
capabilities and limitat... | [
-0.3734329045,
-0.8342722058,
0.3823467195,
0.3023668528,
-0.1560914963,
0.4299522638,
-0.222971499,
-0.561930716,
0.0564764664,
0.2097305804,
-0.2523607612,
-0.5239220858,
-1.1174702644,
0.4230865836,
-0.5675010681,
0.5597125888,
-0.0857368261,
0.1110436246,
0.1416088045,
-0.0... |
http://arxiv.org/abs/2406.12673v1 | http://arxiv.org/pdf/2406.12673v1 | Estimating Knowledge in Large Language Models Without Generating a Single Token | 2024-06-18T14:45:50 | 2024-06-18T14:45:50 | Daniela Gottesman, Mor Geva | cs.CL | cs.CL | To evaluate knowledge in large language models (LLMs), current methods query
the model and then evaluate its generated responses. In this work, we ask
whether evaluation can be done $\textit{before}$ the model has generated any
text. Concretely, is it possible to estimate how knowledgeable a model is about
a certain en... | [
-0.3008795381,
-0.7028089166,
0.9629470706,
0.002974896,
-0.2411068678,
-0.0294322167,
-0.6074746847,
-0.5243496895,
0.3959582448,
0.4432681203,
-0.453805238,
-0.5309284329,
-0.4616382718,
-0.2636595666,
-0.4549657702,
0.8098554611,
0.3465469778,
0.0025819016,
-0.4697254896,
-0... |
http://arxiv.org/abs/2406.12665v1 | http://arxiv.org/pdf/2406.12665v1 | CollabStory: Multi-LLM Collaborative Story Generation and Authorship Analysis | 2024-06-18T14:35:12 | 2024-06-18T14:35:12 | Saranya Venkatraman, Nafis Irtiza Tripto, Dongwon Lee | cs.CL | cs.CL, cs.AI | The rise of unifying frameworks that enable seamless interoperability of
Large Language Models (LLMs) has made LLM-LLM collaboration for open-ended
tasks a possibility. Despite this, there have not been efforts to explore such
collaborative writing. We take the next step beyond human-LLM collaboration to
explore this m... | [
-0.3304794431,
-0.4980128407,
0.6385335922,
0.3241833746,
0.2288163304,
0.5364436507,
-0.333232224,
-0.4289774001,
0.2564256191,
0.6998516917,
-0.8347249031,
-0.7590734363,
-0.4125192165,
0.4224248827,
-0.2371547371,
1.0979633331,
-0.2724430859,
0.0061435406,
-0.9127027988,
-0.... |
http://arxiv.org/abs/2406.12651v1 | http://arxiv.org/pdf/2406.12651v1 | Transforming Surgical Interventions with Embodied Intelligence for Ultrasound Robotics | 2024-06-18T14:22:16 | 2024-06-18T14:22:16 | Huan Xu, Jinlin Wu, Guanglin Cao, Zhen Chen, Zhen Lei, Hongbin Liu | cs.RO | cs.RO, cs.AI, cs.CL, cs.HC | Ultrasonography has revolutionized non-invasive diagnostic methodologies,
significantly enhancing patient outcomes across various medical domains.
Despite its advancements, integrating ultrasound technology with robotic
systems for automated scans presents challenges, including limited command
understanding and dynamic... | [
-0.1078730673,
-0.4787780046,
0.9410204887,
-0.3733218014,
-0.6522044539,
0.354747951,
0.0682993904,
-0.7020223141,
0.6464197636,
0.5220728517,
-0.6981810331,
-0.4877955914,
-0.8547620177,
0.046425391,
-0.4182707071,
0.6574649215,
-0.0129833613,
0.1937786341,
0.0244231708,
-0.4... |
http://arxiv.org/abs/2406.12645v1 | http://arxiv.org/pdf/2406.12645v1 | Evaluating Transparency of Machine Generated Fact Checking Explanations | 2024-06-18T14:13:13 | 2024-06-18T14:13:13 | Rui Xing, Timothy Baldwin, Jey Han Lau | cs.CL | cs.CL, cs.AI | An important factor when it comes to generating fact-checking explanations is
the selection of evidence: intuitively, high-quality explanations can only be
generated given the right evidence. In this work, we investigate the impact of
human-curated vs. machine-selected evidence for explanation generation using
large la... | [
-0.4812930226,
-0.7873802781,
1.0863676071,
-0.059598878,
-0.3522495329,
-0.3275077641,
-0.079092972,
-0.5399173498,
-0.2131010592,
0.6080685258,
-0.570438385,
-0.2390680611,
-0.8475297093,
0.2637118399,
-0.4904571474,
1.1086735725,
-0.128468588,
0.0034823609,
-0.197156623,
0.1... |
http://arxiv.org/abs/2406.12644v2 | http://arxiv.org/pdf/2406.12644v2 | Hierarchical Prompting Taxonomy: A Universal Evaluation Framework for Large Language Models | 2024-06-18T14:12:27 | 2024-06-27T14:32:07 | Devichand Budagam, Sankalp KJ, Ashutosh Kumar, Vinija Jain, Aman Chadha | cs.CL | cs.CL, cs.AI | Assessing the effectiveness of large language models (LLMs) in addressing
diverse tasks is essential for comprehending their strengths and weaknesses.
Conventional evaluation techniques typically apply a single prompting strategy
uniformly across datasets, not considering the varying degrees of task
complexity. We intr... | [
-0.2260364741,
-0.8855093122,
0.4832004011,
0.7687290311,
0.0621118248,
0.1327883154,
-0.2555411756,
-0.0170267131,
-0.3098595142,
-0.0646296218,
-0.6601335406,
-0.5338338614,
-0.3864841461,
0.3723380268,
-0.1052685529,
0.8398020267,
0.0977054909,
-0.0249240492,
-0.3247223794,
... |
http://arxiv.org/abs/2406.12641v1 | http://arxiv.org/pdf/2406.12641v1 | DetectBench: Can Large Language Model Detect and Piece Together Implicit Evidence? | 2024-06-18T14:08:01 | 2024-06-18T14:08:01 | Zhouhong Gu, Lin Zhang, Xiaoxuan Zhu, Jiangjie Chen, Wenhao Huang, Yikai Zhang, Shusen Wang, Zheyu Ye, Yan Gao, Hongwei Feng, Yanghua Xiao | cs.CL | cs.CL | Detecting evidence within the context is a key step in the process of
reasoning task. Evaluating and enhancing the capabilities of LLMs in evidence
detection will strengthen context-based reasoning performance. This paper
proposes a benchmark called DetectBench for verifying the ability to detect and
piece together imp... | [
-0.6732698679,
-0.9271810055,
1.1349081993,
0.0857568979,
-0.3377058506,
-0.2213679403,
-0.2193248272,
-0.7335229516,
-0.1755938828,
0.2242006958,
-0.4982215762,
-0.5828637481,
-0.5639939308,
0.2184406817,
-0.4622789323,
0.5129672885,
-0.0335484706,
0.2084114105,
-0.1489835829,
... |
http://arxiv.org/abs/2406.12639v1 | http://arxiv.org/pdf/2406.12639v1 | Ask-before-Plan: Proactive Language Agents for Real-World Planning | 2024-06-18T14:07:28 | 2024-06-18T14:07:28 | Xuan Zhang, Yang Deng, Zifeng Ren, See-Kiong Ng, Tat-Seng Chua | cs.CL | cs.CL, cs.AI | The evolution of large language models (LLMs) has enhanced the planning
capabilities of language agents in diverse real-world scenarios. Despite these
advancements, the potential of LLM-powered agents to comprehend ambiguous user
instructions for reasoning and decision-making is still under exploration. In
this work, w... | [
-0.6962913871,
-0.9314361215,
0.7567293048,
0.1079809591,
0.0759715214,
-0.0243641995,
-0.3266165555,
-0.3226900995,
0.2004695684,
0.683219254,
-1.2348810434,
-0.2687368095,
-0.0573111847,
-0.2608307302,
-0.0385281444,
0.5692633986,
-0.1332290322,
-0.1965607107,
0.0747379586,
-... |
http://arxiv.org/abs/2406.19413v1 | http://arxiv.org/pdf/2406.19413v1 | Saliency Attention and Semantic Similarity-Driven Adversarial Perturbation | 2024-06-18T14:07:27 | 2024-06-18T14:07:27 | Hetvi Waghela, Jaydip Sen, Sneha Rakshit | cs.CR | cs.CR, cs.CL, cs.LG | In this paper, we introduce an enhanced textual adversarial attack method,
known as Saliency Attention and Semantic Similarity driven adversarial
Perturbation (SASSP). The proposed scheme is designed to improve the
effectiveness of contextual perturbations by integrating saliency, attention,
and semantic similarity. Tr... | [
-0.5396862626,
-1.0036040545,
0.3948235512,
0.3807190657,
-0.4655580521,
-0.2411980033,
-0.1106231883,
-0.0742113069,
0.3596940935,
0.9708011746,
-0.6444030404,
-0.4990199506,
-0.7622129321,
0.1279753149,
-0.7178217173,
1.0459438562,
0.3674233556,
-0.0178475659,
-0.1699844748,
... |
http://arxiv.org/abs/2406.12629v1 | http://arxiv.org/pdf/2406.12629v1 | SeTAR: Out-of-Distribution Detection with Selective Low-Rank Approximation | 2024-06-18T13:55:13 | 2024-06-18T13:55:13 | Yixia Li, Boya Xiong, Guanhua Chen, Yun Chen | cs.CL | cs.CL, cs.AI, cs.CV | Out-of-distribution (OOD) detection is crucial for the safe deployment of
neural networks. Existing CLIP-based approaches perform OOD detection by
devising novel scoring functions or sophisticated fine-tuning methods. In this
work, we propose SeTAR, a novel, training-free OOD detection method that
leverages selective l... | [
-0.4905878901,
-0.26393646,
0.1318756193,
-0.138191849,
-0.5691185594,
-0.0469640158,
-0.0215772036,
-0.3863711059,
0.2384888381,
0.6601784229,
-0.2406497598,
-0.1343354732,
-0.5500238538,
-0.2880438566,
-0.3655748665,
0.3709525466,
0.0157802049,
-0.1778510362,
0.2071564645,
-0... |
http://arxiv.org/abs/2406.12624v2 | http://arxiv.org/pdf/2406.12624v2 | Judging the Judges: Evaluating Alignment and Vulnerabilities in LLMs-as-Judges | 2024-06-18T13:49:54 | 2024-07-02T02:57:01 | Aman Singh Thakur, Kartik Choudhary, Venkat Srinik Ramayapally, Sankaran Vaidyanathan, Dieuwke Hupkes | cs.CL | cs.CL, cs.AI | Offering a promising solution to the scalability challenges associated with
human evaluation, the LLM-as-a-judge paradigm is rapidly gaining traction as an
approach to evaluating large language models (LLMs). However, there are still
many open questions about the strengths and weaknesses of this paradigm, and
what pote... | [
-0.4022143781,
-0.4634779394,
0.919593215,
-0.0532502048,
-0.2791363895,
0.1809386313,
-0.3009948432,
-0.5664590001,
0.0302110873,
0.5663204193,
-0.073107399,
-0.453922838,
-0.8601541519,
0.017729722,
-0.5193685889,
0.8694646955,
0.0367748663,
0.525552392,
-0.4604495764,
-0.081... |
http://arxiv.org/abs/2406.12621v1 | http://arxiv.org/pdf/2406.12621v1 | Growing Trees on Sounds: Assessing Strategies for End-to-End Dependency Parsing of Speech | 2024-06-18T13:46:10 | 2024-06-18T13:46:10 | Adrien Pupier, Maximin Coavoux, Jérôme Goulian, Benjamin Lecouteux | cs.CL | cs.CL | Direct dependency parsing of the speech signal -- as opposed to parsing
speech transcriptions -- has recently been proposed as a task (Pupier et al.
2022), as a way of incorporating prosodic information in the parsing system and
bypassing the limitations of a pipeline approach that would consist of using
first an Autom... | [
-0.3847923279,
-0.8985239863,
0.5272586942,
0.5297572017,
-0.2098225057,
0.3927762508,
-0.1203860492,
-0.876223743,
0.3182921708,
0.2958199382,
-0.43977198,
-0.2918077111,
-0.6648811698,
0.369361639,
-0.594717145,
0.4369631708,
0.2915837467,
-0.0418670587,
-0.1194767505,
-0.096... |
http://arxiv.org/abs/2406.12620v2 | http://arxiv.org/pdf/2406.12620v2 | What Makes Two Language Models Think Alike? | 2024-06-18T13:45:50 | 2024-06-24T06:34:35 | Jeanne Salle, Louis Jalouzot, Nur Lan, Emmanuel Chemla, Yair Lakretz | cs.CL | cs.CL | Do architectural differences significantly affect the way models represent
and process language? We propose a new approach, based on metric-learning
encoding models (MLEMs), as a first step to answer this question. The approach
provides a feature-based comparison of how any two layers of any two models
represent lingui... | [
-0.5500035286,
-0.5519962907,
0.5221884847,
0.1040695235,
-0.0766602457,
0.0802595764,
-0.1738014519,
-0.6387578249,
0.0568908378,
0.5025021434,
-0.4122287333,
-0.3371106684,
-0.7197733521,
-0.0118674273,
-0.1226105541,
0.6802257299,
0.0319514759,
0.4284747541,
-0.2095112205,
-... |
http://arxiv.org/abs/2406.12618v1 | http://arxiv.org/pdf/2406.12618v1 | From Insights to Actions: The Impact of Interpretability and Analysis Research on NLP | 2024-06-18T13:45:07 | 2024-06-18T13:45:07 | Marius Mosbach, Vagrant Gautam, Tomás Vergara-Browne, Dietrich Klakow, Mor Geva | cs.CL | cs.CL | Interpretability and analysis (IA) research is a growing subfield within NLP
with the goal of developing a deeper understanding of the behavior or inner
workings of NLP systems and methods. Despite growing interest in the subfield,
a commonly voiced criticism is that it lacks actionable insights and therefore
has littl... | [
-0.1553649604,
-0.5952814221,
0.3813036084,
0.4213907719,
0.103295289,
0.0695302039,
0.1248979792,
-0.9703173637,
0.6823058724,
0.3012065589,
-0.1376523972,
-0.1569910049,
-0.906052649,
0.2562153041,
-0.69030267,
1.0308494568,
0.1914077848,
-0.1492271572,
-0.2328420132,
-0.5275... |
http://arxiv.org/abs/2406.12614v1 | http://arxiv.org/pdf/2406.12614v1 | EUvsDisinfo: a Dataset for Multilingual Detection of Pro-Kremlin Disinformation in News Articles | 2024-06-18T13:43:22 | 2024-06-18T13:43:22 | João A. Leite, Olesya Razuvayevskaya, Kalina Bontcheva, Carolina Scarton | cs.CL | cs.CL, cs.LG | This work introduces EUvsDisinfo, a multilingual dataset of trustworthy and
disinformation articles related to pro-Kremlin themes. It is sourced directly
from the debunk articles written by experts leading the EUvsDisinfo project.
Our dataset is the largest to-date resource in terms of the overall number of
articles an... | [
-0.2411090285,
-0.5860531926,
0.3745902181,
0.1884152442,
-0.542365551,
0.458247602,
-0.2153115124,
-0.1956316829,
0.0619735979,
0.1222857311,
-0.5515916944,
-0.7436254621,
-0.690843761,
0.2729120553,
-0.6161885858,
0.9930009246,
0.001955145,
0.5694125295,
-0.2335569263,
0.0070... |
http://arxiv.org/abs/2406.12611v1 | http://arxiv.org/pdf/2406.12611v1 | Rapid Language Adaptation for Multilingual E2E Speech Recognition Using Encoder Prompting | 2024-06-18T13:38:58 | 2024-06-18T13:38:58 | Yosuke Kashiwagi, Hayato Futami, Emiru Tsunoo, Siddhant Arora, Shinji Watanabe | cs.SD | cs.SD, cs.CL, eess.AS | End-to-end multilingual speech recognition models handle multiple languages
through a single model, often incorporating language identification to
automatically detect the language of incoming speech. Since the common scenario
is where the language is already known, these models can perform as
language-specific by usin... | [
-0.2376678437,
-0.665004909,
0.2915592492,
0.3116911054,
-0.1537895054,
-0.1140994057,
-0.4085683525,
-0.238014102,
0.0750024244,
0.0675595328,
-0.5461761951,
-0.166476056,
-0.3533764184,
0.5942953825,
-0.0902385712,
0.6571859717,
-0.0686077401,
-0.0830007717,
0.0441578887,
-0.... |
http://arxiv.org/abs/2406.12608v1 | http://arxiv.org/pdf/2406.12608v1 | Bridging Local Details and Global Context in Text-Attributed Graphs | 2024-06-18T13:35:25 | 2024-06-18T13:35:25 | Yaoke Wang, Yun Zhu, Wenqiao Zhang, Yueting Zhuang, Yunfei Li, Siliang Tang | cs.CL | cs.CL, cs.AI | Representation learning on text-attributed graphs (TAGs) is vital for
real-world applications, as they combine semantic textual and contextual
structural information. Research in this field generally consist of two main
perspectives: local-level encoding and global-level aggregating, respectively
refer to textual node ... | [
-0.1953683943,
-0.760584712,
0.417222321,
-0.0271999296,
-0.173615247,
0.3230163455,
0.0815111771,
-0.7457267642,
0.2199625075,
0.2636566162,
-0.256883055,
-0.6812074184,
-0.4698831737,
-0.1345130652,
-0.2490981966,
0.9582518935,
-0.4166409671,
-0.2354176491,
-0.6191195846,
-0.... |
http://arxiv.org/abs/2406.12606v1 | http://arxiv.org/pdf/2406.12606v1 | Low-Redundant Optimization for Large Language Model Alignment | 2024-06-18T13:34:40 | 2024-06-18T13:34:40 | Zhipeng Chen, Kun Zhou, Wayne Xin Zhao, Jingyuan Wang, Ji-Rong Wen | cs.CL | cs.CL | Large language models (LLMs) are still struggling in aligning with human
preference in complex tasks and scenarios. They are prone to overfit into the
unexpected patterns or superficial styles in the training data. We conduct an
empirical study that only selects the top-10\% most updated parameters in LLMs
for alignmen... | [
-0.0586744808,
-0.4532680511,
0.5317854881,
0.0237571318,
-0.2087601423,
-0.1247642711,
-0.4279031157,
-0.2547543943,
0.3646464944,
0.2583502531,
-0.5716133714,
-0.483589083,
-0.5227764845,
-0.0922972485,
-0.1840098649,
0.8833035827,
0.0396043845,
0.1361831576,
-0.4445141852,
-... |
http://arxiv.org/abs/2406.12593v1 | http://arxiv.org/pdf/2406.12593v1 | PromptDSI: Prompt-based Rehearsal-free Instance-wise Incremental Learning for Document Retrieval | 2024-06-18T13:25:18 | 2024-06-18T13:25:18 | Tuan-Luc Huynh, Thuy-Trang Vu, Weiqing Wang, Yinwei Wei, Trung Le, Dragan Gasevic, Yuan-Fang Li, Thanh-Toan Do | cs.IR | cs.IR, cs.AI, cs.CL, cs.LG | Differentiable Search Index (DSI) utilizes Pre-trained Language Models (PLMs)
for efficient document retrieval without relying on external indexes. However,
DSIs need full re-training to handle updates in dynamic corpora, causing
significant computational inefficiencies. We introduce PromptDSI, a
rehearsal-free, prompt... | [
-0.2088148147,
-0.3264585733,
0.7427133918,
0.1746459305,
-0.2850278914,
0.1543818414,
-0.0922737569,
0.4128566086,
0.2638860643,
0.0246660262,
-0.9278348684,
-0.578050673,
-0.2582965791,
0.2500678003,
-0.2180930078,
0.5682507753,
0.045296248,
0.1349453032,
-0.4193306267,
-0.18... |
http://arxiv.org/abs/2406.12585v1 | http://arxiv.org/pdf/2406.12585v1 | Breaking the Ceiling of the LLM Community by Treating Token Generation as a Classification for Ensembling | 2024-06-18T13:17:26 | 2024-06-18T13:17:26 | Yao-Ching Yu, Chun-Chih Kuo, Ziqi Ye, Yu-Cheng Chang, Yueh-Se Li | cs.CL | cs.CL, cs.AI | Ensembling multiple models has always been an effective approach to push the
limits of existing performance and is widely used in classification tasks by
simply averaging the classification probability vectors from multiple
classifiers to achieve better accuracy. However, in the thriving open-source
Large Language Mode... | [
-0.1935859174,
-0.7599255443,
0.5228821635,
-0.1427152753,
-0.4441413879,
0.4067093134,
-0.175161913,
-0.4006612897,
0.1910074055,
0.5172325969,
-0.2934700847,
-0.6360454559,
-0.7663709521,
-0.0812058151,
-0.4465983808,
0.814101696,
0.0591647141,
-0.2356355935,
-0.6182310581,
-... |
http://arxiv.org/abs/2406.12572v2 | http://arxiv.org/pdf/2406.12572v2 | Mathador-LM: A Dynamic Benchmark for Mathematical Reasoning on Large Language Models | 2024-06-18T13:02:12 | 2024-06-19T12:28:10 | Eldar Kurtic, Amir Moeini, Dan Alistarh | cs.CL | cs.CL, cs.AI, cs.LG, I.2.7 | We introduce Mathador-LM, a new benchmark for evaluating the mathematical
reasoning on large language models (LLMs), combining ruleset interpretation,
planning, and problem-solving. This benchmark is inspired by the Mathador game,
where the objective is to reach a target number using basic arithmetic
operations on a gi... | [
-0.1791445762,
-0.7130427957,
0.5108408332,
0.5444877744,
-0.159828946,
-0.0224434528,
-0.3570322692,
-0.360563457,
-0.0552306697,
0.3270592988,
-0.6917488575,
-0.6345539093,
-0.5131574273,
-0.1554769427,
-0.3105248809,
0.643148005,
-0.3773795068,
0.2048783153,
-0.375400722,
-0... |
http://arxiv.org/abs/2406.15490v1 | http://arxiv.org/pdf/2406.15490v1 | Causal Discovery Inspired Unsupervised Domain Adaptation for Emotion-Cause Pair Extraction | 2024-06-18T13:01:30 | 2024-06-18T13:01:30 | Yuncheng Hua, Yujin Huang, Shuo Huang, Tao Feng, Lizhen Qu, Chris Bain, Richard Bassed, Gholamreza Haffari | cs.CL | cs.CL, cs.AI, cs.LG, I.2.4 | This paper tackles the task of emotion-cause pair extraction in the
unsupervised domain adaptation setting. The problem is challenging as the
distributions of the events causing emotions in target domains are dramatically
different than those in source domains, despite the distributions of emotional
expressions between... | [
-0.2897671163,
-0.4982588291,
0.4709902108,
0.3568603992,
-0.2387446314,
-0.1555330902,
-0.0605426393,
-0.3467722237,
0.071581617,
0.1161167622,
-0.7026402354,
-0.3740353584,
-0.529027164,
-0.2320250273,
-0.1338220388,
0.8962652683,
-0.1693477482,
-0.0437509604,
-0.096457243,
-... |
http://arxiv.org/abs/2406.12570v1 | http://arxiv.org/pdf/2406.12570v1 | Applying Ensemble Methods to Model-Agnostic Machine-Generated Text Detection | 2024-06-18T12:58:01 | 2024-06-18T12:58:01 | Ivan Ong, Boon King Quek | cs.CL | cs.CL | In this paper, we study the problem of detecting machine-generated text when
the large language model (LLM) it is possibly derived from is unknown. We do so
by apply ensembling methods to the outputs from DetectGPT classifiers (Mitchell
et al. 2023), a zero-shot model for machine-generated text detection which is
highl... | [
-0.1879162788,
-0.9281965494,
0.7762900591,
-0.2408640832,
-0.1348323375,
0.0732550398,
-0.1265824288,
-0.6867105365,
0.0724245459,
0.4257751703,
-0.3316725492,
-0.5142610073,
-0.7342330813,
0.2510153949,
-0.8224565983,
0.8865618706,
-0.0966893435,
-0.4453864694,
0.1286989748,
... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.