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/2406.16536v1 | http://arxiv.org/pdf/2406.16536v1 | C-LLM: Learn to Check Chinese Spelling Errors Character by Character | 2024-06-24T11:16:31 | 2024-06-24T11:16:31 | Kunting Li, Yong Hu, Liang He, Fandong Meng, Jie Zhou | cs.CL | cs.CL | Chinese Spell Checking (CSC) aims to detect and correct spelling errors in
sentences. Despite Large Language Models (LLMs) exhibit robust capabilities and
are widely applied in various tasks, their performance on CSC is often
unsatisfactory. We find that LLMs fail to meet the Chinese character-level
constraints of the ... | [
0.4049620032,
-0.607468009,
0.5466833115,
0.2996390462,
-0.1401839852,
0.3841864765,
-0.2044908255,
-0.5043369532,
0.1442667097,
0.491854459,
-0.5534549952,
-0.9115970731,
-0.538837254,
0.3130769432,
-0.3060038984,
0.3982437849,
-0.2238663435,
0.4619745612,
-0.0684817731,
0.215... |
http://arxiv.org/abs/2406.16535v1 | http://arxiv.org/pdf/2406.16535v1 | Token-based Decision Criteria Are Suboptimal in In-context Learning | 2024-06-24T11:16:26 | 2024-06-24T11:16:26 | Hakaze Cho, Yoshihiro Sakai, Mariko Kato, Kenshiro Tanaka, Akira Ishii, Naoya Inoue | cs.CL | cs.CL, cs.AI, cs.LG | In-Context Learning (ICL) typically utilizes classification criteria from
probabilities of manually selected label tokens. However, we argue that such
token-based classification criteria lead to suboptimal decision boundaries,
despite delicate calibrations through translation and constrained rotation. To
address this p... | [
-0.4588796198,
-0.1383791715,
0.4273662865,
-0.4095616639,
-0.3641046882,
0.315364778,
-0.2364786267,
-0.2783195972,
-0.0176988374,
0.3367698193,
-0.551869154,
-0.7305810452,
-0.2769970596,
-0.156815514,
0.0275961813,
1.121625185,
0.068917729,
0.2105044276,
-0.557964921,
-0.500... |
http://arxiv.org/abs/2406.16529v1 | http://arxiv.org/pdf/2406.16529v1 | Towards Better Graph-based Cross-document Relation Extraction via Non-bridge Entity Enhancement and Prediction Debiasing | 2024-06-24T11:08:28 | 2024-06-24T11:08:28 | Hao Yue, Shaopeng Lai, Chengyi Yang, Liang Zhang, Junfeng Yao, Jinsong Su | cs.CL | cs.CL | Cross-document Relation Extraction aims to predict the relation between
target entities located in different documents. In this regard, the dominant
models commonly retain useful information for relation prediction via bridge
entities, which allows the model to elaborately capture the intrinsic
interdependence between ... | [
-0.5871431828,
-0.7200073004,
0.4246395826,
-0.2238531858,
-0.3213609457,
0.1079995632,
-0.005813757,
-0.7010891438,
0.4082594514,
0.5511006117,
-0.4288022518,
-0.7156952024,
-0.899448812,
0.088326335,
-0.344552815,
1.0043151379,
-0.0962805003,
-0.0014738848,
-0.2282507718,
-0.... |
http://arxiv.org/abs/2406.16528v1 | http://arxiv.org/pdf/2406.16528v1 | Evaluating the Ability of Large Language Models to Reason about Cardinal Directions | 2024-06-24T11:07:01 | 2024-06-24T11:07:01 | Anthony G Cohn, Robert E Blackwell | cs.CL | cs.CL | We investigate the abilities of a representative set of Large language Models
(LLMs) to reason about cardinal directions (CDs). To do so, we create two
datasets: the first, co-created with ChatGPT, focuses largely on recall of
world knowledge about CDs; the second is generated from a set of templates,
comprehensively t... | [
-0.1860888749,
-0.6592460275,
0.8814703822,
0.560560286,
-0.2915278673,
-0.3144777119,
-0.7092846036,
-0.1240380108,
-0.1120300665,
0.6005879045,
-0.5295235515,
-0.7771674991,
-0.4339205325,
0.1374829262,
-0.6058472395,
0.8781075478,
0.0720116124,
0.0229375958,
-0.3648762703,
-... |
http://arxiv.org/abs/2406.16527v1 | http://arxiv.org/pdf/2406.16527v1 | SyROCCo: Enhancing Systematic Reviews using Machine Learning | 2024-06-24T11:04:43 | 2024-06-24T11:04:43 | Zheng Fang, Miguel Arana-Catania, Felix-Anselm van Lier, Juliana Outes Velarde, Harry Bregazzi, Mara Airoldi, Eleanor Carter, Rob Procter | cs.CL | cs.CL, cs.CY, cs.DL, cs.LG | The sheer number of research outputs published every year makes systematic
reviewing increasingly time- and resource-intensive. This paper explores the
use of machine learning techniques to help navigate the systematic review
process. ML has previously been used to reliably 'screen' articles for review -
that is, ident... | [
-0.5538677573,
-0.3005287051,
0.4666953385,
0.1337329447,
-0.3209890127,
-0.1057728231,
0.0723176077,
-0.294200778,
0.1168595254,
0.6574379802,
-0.6375038624,
-0.6925716996,
-0.7425784469,
0.3498868048,
-0.5426294208,
1.1735363007,
-0.1912659258,
0.3321684003,
0.0187251382,
-0.... |
http://arxiv.org/abs/2406.16524v1 | http://arxiv.org/pdf/2406.16524v1 | The Privileged Students: On the Value of Initialization in Multilingual Knowledge Distillation | 2024-06-24T10:59:26 | 2024-06-24T10:59:26 | Haryo Akbarianto Wibowo, Thamar Solorio, Alham Fikri Aji | cs.CL | cs.CL, 68T50 | Knowledge distillation (KD) has proven to be a successful strategy to improve
the performance of a smaller model in many NLP tasks. However, most of the work
in KD only explores monolingual scenarios. In this paper, we investigate the
value of KD in multilingual settings. We find the significance of KD and model
initia... | [
-0.1666941941,
-0.2170035094,
0.3161325157,
0.190462932,
-0.3966103196,
0.4304871261,
-0.3388276398,
-0.2913075686,
-0.1232347712,
0.1384268403,
0.1603094488,
-0.132436797,
-0.8803573847,
0.0456326902,
-0.1745815128,
1.406773448,
0.0577161573,
0.3109827638,
-0.3935609758,
-0.14... |
http://arxiv.org/abs/2406.16521v1 | http://arxiv.org/pdf/2406.16521v1 | Carrot and Stick: Inducing Self-Motivation with Positive & Negative Feedback | 2024-06-24T10:55:31 | 2024-06-24T10:55:31 | Jimin Sohn, Jeihee Cho, Junyong Lee, Songmu Heo, Ji-Eun Han, David R. Mortensen | cs.CL | cs.CL, cs.AI | Positive thinking is thought to be an important component of self-motivation
in various practical fields such as education and the workplace. Previous work,
including sentiment transfer and positive reframing, has focused on the
positive side of language. However, self-motivation that drives people to reach
their goals... | [
-0.225688681,
-0.6254960895,
0.4112913311,
0.5443156362,
-0.1471016854,
0.1677452922,
-0.0494988635,
-0.2975206375,
0.5217558742,
0.1394105703,
-0.7312162519,
-0.3149552047,
-0.3865348399,
0.245262146,
-0.2606198788,
1.1271796227,
0.1243582517,
0.3194838464,
-0.2430063337,
-0.0... |
http://arxiv.org/abs/2406.16508v1 | http://arxiv.org/pdf/2406.16508v1 | Large Vocabulary Size Improves Large Language Models | 2024-06-24T10:27:07 | 2024-06-24T10:27:07 | Sho Takase, Ryokan Ri, Shun Kiyono, Takuya Kato | cs.CL | cs.CL | This paper empirically investigates the relationship between subword
vocabulary size and the performance of large language models (LLMs) to provide
insights on how to define the vocabulary size. Experimental results show that
larger vocabulary sizes lead to better performance in LLMs. Moreover, we
consider a continual ... | [
-0.5317494869,
-0.6927598715,
0.489122957,
0.1547150165,
-0.5630888939,
0.2079406232,
-0.6657736897,
-0.6949591637,
0.2975798547,
0.4438105822,
-0.073895961,
-0.0863280371,
-0.8965512514,
0.1563900113,
-0.5747603178,
0.6523209214,
0.0293402392,
0.3290278912,
-0.4258314967,
-0.0... |
http://arxiv.org/abs/2406.16495v2 | http://arxiv.org/pdf/2406.16495v2 | OTCE: Hybrid SSM and Attention with Cross Domain Mixture of Experts to construct Observer-Thinker-Conceiver-Expresser | 2024-06-24T10:05:23 | 2024-06-25T02:20:14 | Jingze Shi, Ting Xie, Bingheng Wu, Chunjun Zheng, Kai Wang | cs.CL | cs.CL, cs.AI | Recent research has shown that combining Mamba with Transformer architecture,
which has selective state space and quadratic self-attention mechanism,
outperforms using Mamba or Transformer architecture alone in language modeling
tasks. The quadratic self-attention mechanism effectively alleviates the
shortcomings of se... | [
-0.5949522257,
-0.5424191952,
0.6656930447,
-0.081563592,
-0.0545400716,
0.4163135886,
-0.1102834344,
-0.7279782891,
0.5073285699,
0.2717137039,
-0.4445546567,
-0.2114659846,
-0.5729106665,
-0.036683768,
-0.2089474201,
0.426109314,
-0.1077877656,
-0.0852430463,
0.0166265387,
-0... |
http://arxiv.org/abs/2406.16490v1 | http://arxiv.org/pdf/2406.16490v1 | eagerlearners at SemEval2024 Task 5: The Legal Argument Reasoning Task in Civil Procedure | 2024-06-24T09:57:44 | 2024-06-24T09:57:44 | Hoorieh Sabzevari, Mohammadmostafa Rostamkhani, Sauleh Eetemadi | cs.CL | cs.CL | This study investigates the performance of the zero-shot method in
classifying data using three large language models, alongside two models with
large input token sizes and the two pre-trained models on legal data. Our main
dataset comes from the domain of U.S. civil procedure. It includes summaries of
legal cases, spe... | [
0.0470282659,
-0.7603878379,
0.926830411,
0.0371436439,
-0.4303082526,
-0.3985082209,
-0.0472502187,
-0.2764877379,
-0.3094845116,
0.3647347689,
-0.1255057901,
-0.7363890409,
-0.5890854597,
-0.0420183353,
-0.5915017128,
0.8803963065,
-0.0221347958,
0.5123562813,
-0.307351023,
-... |
http://arxiv.org/abs/2406.16489v1 | http://arxiv.org/pdf/2406.16489v1 | Deepfake tweets automatic detection | 2024-06-24T09:55:31 | 2024-06-24T09:55:31 | Adam Frej, Adrian Kaminski, Piotr Marciniak, Szymon Szmajdzinski, Soveatin Kuntur, Anna Wroblewska | cs.CL | cs.CL | This study addresses the critical challenge of detecting DeepFake tweets by
leveraging advanced natural language processing (NLP) techniques to distinguish
between genuine and AI-generated texts. Given the increasing prevalence of
misinformation, our research utilizes the TweepFake dataset to train and
evaluate various... | [
-0.4262025356,
-0.9249043465,
0.349262923,
0.2731662691,
-0.1286641806,
0.5979166627,
0.2183124125,
-0.839689672,
0.0689034462,
0.5427360535,
-0.8623415828,
-0.3434500098,
-0.7608836889,
0.2540989518,
-0.658000946,
0.6533998251,
-0.0050505931,
-0.3378114998,
0.5104268789,
0.178... |
http://arxiv.org/abs/2406.16478v1 | http://arxiv.org/pdf/2406.16478v1 | EMMI -- Empathic Multimodal Motivational Interviews Dataset: Analyses and Annotations | 2024-06-24T09:32:28 | 2024-06-24T09:32:28 | Lucie Galland, Catherine Pelachaud, Florian Pecune | cs.CL | cs.CL | The study of multimodal interaction in therapy can yield a comprehensive
understanding of therapist and patient behavior that can be used to develop a
multimodal virtual agent supporting therapy. This investigation aims to uncover
how therapists skillfully blend therapy's task goal (employing classical steps
of Motivat... | [
-0.389740169,
-0.9136160016,
0.8062000275,
0.2541399598,
0.1028250381,
0.1804468781,
-0.1149072945,
-0.2615431249,
0.7598219514,
0.2385548651,
-0.638156414,
-0.325846523,
-0.5697928071,
0.3707298934,
0.2309278101,
0.8633877635,
-0.1328860968,
0.4475604892,
-0.47806409,
-0.27731... |
http://arxiv.org/abs/2406.16477v1 | http://arxiv.org/pdf/2406.16477v1 | DaLPSR: Leverage Degradation-Aligned Language Prompt for Real-World Image Super-Resolution | 2024-06-24T09:30:36 | 2024-06-24T09:30:36 | Aiwen Jiang, Zhi Wei, Long Peng, Feiqiang Liu, Wenbo Li, Mingwen Wang | cs.CV | cs.CV, cs.CL | Image super-resolution pursuits reconstructing high-fidelity high-resolution
counterpart for low-resolution image. In recent years, diffusion-based models
have garnered significant attention due to their capabilities with rich prior
knowledge. The success of diffusion models based on general text prompts has
validated ... | [
-0.4509768784,
-1.0145279169,
0.742092669,
-0.1544899344,
-0.2640995085,
-0.0883866176,
-0.0463275649,
-0.3096894026,
-0.0604200363,
0.4244038463,
-0.4693065584,
-0.2402279973,
-0.5717126727,
0.1831538826,
-0.4206094742,
0.8261635303,
0.1067182794,
0.2526614368,
-0.2564449608,
... |
http://arxiv.org/abs/2406.16469v1 | http://arxiv.org/pdf/2406.16469v1 | Evaluating Visual and Cultural Interpretation: The K-Viscuit Benchmark with Human-VLM Collaboration | 2024-06-24T09:18:15 | 2024-06-24T09:18:15 | Yujin Baek, ChaeHun Park, Jaeseok Kim, Yu-Jung Heo, Du-Seong Chang, Jaegul Choo | cs.CL | cs.CL, cs.CV | To create culturally inclusive vision-language models (VLMs), the foremost
requirement is developing a test benchmark that can diagnose the models'
ability to respond to questions reflecting cultural elements. This paper
addresses the necessity for such benchmarks, noting that existing research has
relied on human anno... | [
-0.4883060753,
0.0321260206,
0.4185653329,
0.1842241585,
-0.3238168657,
0.1973466724,
-0.5015375614,
-0.9324942231,
-0.05842885,
0.6488110423,
-0.4649004936,
-0.6711822748,
-0.2790798843,
0.1309905946,
-0.0791781098,
0.4126355648,
-0.2013300955,
0.3931531906,
-0.3406446874,
-0.... |
http://arxiv.org/abs/2406.16464v2 | http://arxiv.org/pdf/2406.16464v2 | InterCLIP-MEP: Interactive CLIP and Memory-Enhanced Predictor for Multi-modal Sarcasm Detection | 2024-06-24T09:13:42 | 2024-06-26T05:40:16 | Junjie Chen, Subin Huang | cs.CL | cs.CL, cs.AI, cs.CV | The prevalence of sarcasm in social media, conveyed through text-image
combinations, presents significant challenges for sentiment analysis and
intention mining. Current multi-modal sarcasm detection methods have been
proven to struggle with biases from spurious cues, leading to a superficial
understanding of the compl... | [
-0.595490396,
-0.4419465959,
0.2460935563,
0.6320976615,
-0.3963934779,
0.3567996919,
-0.1492901146,
-0.2440871894,
0.6370689869,
-0.1798355877,
-0.8065298796,
-0.1861898005,
-0.5863320231,
-0.0626781732,
0.0032667331,
0.8893470168,
0.2109295428,
0.3458064795,
-0.138115868,
-0.... |
http://arxiv.org/abs/2406.16450v1 | http://arxiv.org/pdf/2406.16450v1 | Building on Efficient Foundations: Effectively Training LLMs with Structured Feedforward Layers | 2024-06-24T08:43:21 | 2024-06-24T08:43:21 | Xiuying Wei, Skander Moalla, Razvan Pascanu, Caglar Gulcehre | cs.CL | cs.CL | State-of-the-art results in large language models (LLMs) often rely on scale,
which becomes computationally expensive. This has sparked a research agenda to
reduce these models' parameter count and computational costs without
significantly impacting their performance. Our study focuses on
transformer-based LLMs, specif... | [
-0.265858531,
-0.4308564663,
0.29591766,
0.647536397,
0.0353952162,
0.0621005185,
-0.3171465993,
-0.2919102013,
0.1747171432,
0.4811019003,
-0.6607464552,
-0.0491408557,
-0.5088343024,
0.1045606211,
-0.5992634296,
0.6879345775,
0.0966151208,
0.1450839192,
0.0480310172,
0.035008... |
http://arxiv.org/abs/2406.16441v1 | http://arxiv.org/pdf/2406.16441v1 | UniCoder: Scaling Code Large Language Model via Universal Code | 2024-06-24T08:32:48 | 2024-06-24T08:32:48 | Tao Sun, Linzheng Chai, Jian Yang, Yuwei Yin, Hongcheng Guo, Jiaheng Liu, Bing Wang, Liqun Yang, Zhoujun Li | cs.CL | cs.CL | Intermediate reasoning or acting steps have successfully improved large
language models (LLMs) for handling various downstream natural language
processing (NLP) tasks. When applying LLMs for code generation, recent works
mainly focus on directing the models to articulate intermediate
natural-language reasoning steps, a... | [
-0.333091557,
-0.6289067864,
0.2793225944,
0.2678126991,
-0.0861495733,
-0.1129259169,
-0.5402363539,
-0.3402642012,
-0.0693720281,
0.6213051081,
-0.3905515671,
-0.2763023078,
-0.4419188201,
0.413061887,
-0.2086622864,
0.9759841561,
-0.0448724255,
0.0907560363,
0.0060582957,
-0... |
http://arxiv.org/abs/2406.16416v1 | http://arxiv.org/pdf/2406.16416v1 | Multilingual Knowledge Editing with Language-Agnostic Factual Neurons | 2024-06-24T08:06:56 | 2024-06-24T08:06:56 | Xue zhang, Yunlong Liang, Fandong Meng, Songming Zhang, Yufeng Chen, Jinan Xu, Jie Zhou | cs.CL | cs.CL | Multilingual knowledge editing (MKE) aims to simultaneously revise factual
knowledge across multilingual languages within large language models (LLMs).
However, most existing MKE methods just adapt existing monolingual editing
methods to multilingual scenarios, overlooking the deep semantic connections of
the same fact... | [
-0.2892818451,
-0.6892231703,
0.6693358421,
0.2439615577,
-0.1325726211,
0.2477608919,
-0.4497102797,
-0.4308881462,
0.4663028121,
0.2919952273,
-0.4881270826,
-0.5480476022,
-0.4169722497,
0.1729032546,
0.1029691771,
0.7703714371,
0.0431664735,
0.0562758557,
-0.249940142,
-0.6... |
http://arxiv.org/abs/2406.16408v1 | http://arxiv.org/pdf/2406.16408v1 | A Symmetry Property of Christoffel Words | 2024-06-24T08:05:21 | 2024-06-24T08:05:21 | Yan Lanciault, Christophe Reutenauer | math.CO | math.CO, cs.CL | Motivated by the theory of trapezoidal words, whose sequences of cardinality
of factors by length are symmetric, we introduce a bivariate variant of this
symmetry. We show that this symmetry characterizes Christoffel words, and
establish other related results. | [
-0.2281792462,
-0.1032673717,
0.4511584938,
0.4304001331,
-0.5478321314,
0.2910459638,
-0.2625836432,
-0.03091993,
0.5393846631,
0.6062273383,
-0.5904193521,
-0.6020567417,
-0.7579712868,
0.2050333619,
-0.4954398572,
0.9757781029,
-0.4173689187,
0.6408897638,
0.2884353101,
-0.6... |
http://arxiv.org/abs/2406.16382v1 | http://arxiv.org/pdf/2406.16382v1 | UNO Arena for Evaluating Sequential Decision-Making Capability of Large Language Models | 2024-06-24T07:47:34 | 2024-06-24T07:47:34 | Zhanyue Qin, Haochuan Wang, Deyuan Liu, Ziyang Song, Cunhang Fan, Zhao Lv, Jinlin Wu, Zhen Lei, Zhiying Tu, Dianhui Chu, Xiaoyan Yu, Dianbo Sui | cs.CL | cs.CL | Sequential decision-making refers to algorithms that take into account the
dynamics of the environment, where early decisions affect subsequent decisions.
With large language models (LLMs) demonstrating powerful capabilities between
tasks, we can't help but ask: Can Current LLMs Effectively Make Sequential
Decisions? I... | [
-0.3749651313,
-0.8622558117,
0.6892254949,
0.5151488185,
-0.2210862786,
0.292368412,
-0.3182432055,
-0.3020095527,
-0.073887229,
0.3151672482,
-0.3018829823,
-0.196091339,
-0.2939955592,
0.4081982374,
-0.0861290842,
0.8436831236,
-0.061441686,
0.1414786875,
-0.2787012458,
-0.1... |
http://arxiv.org/abs/2406.16377v1 | http://arxiv.org/pdf/2406.16377v1 | On the Transformations across Reward Model, Parameter Update, and In-Context Prompt | 2024-06-24T07:42:32 | 2024-06-24T07:42:32 | Deng Cai, Huayang Li, Tingchen Fu, Siheng Li, Weiwen Xu, Shuaiyi Li, Bowen Cao, Zhisong Zhang, Xinting Huang, Leyang Cui, Yan Wang, Lemao Liu, Taro Watanabe, Shuming Shi | cs.CL | cs.CL, cs.AI | Despite the general capabilities of pre-trained large language models (LLMs),
they still need further adaptation to better serve practical applications. In
this paper, we demonstrate the interchangeability of three popular and distinct
adaptation tools: parameter updating, reward modeling, and in-context
prompting. Thi... | [
-0.2415413558,
-0.3282194138,
0.9752231836,
0.4184328318,
-0.1124702096,
-0.0877843276,
-0.5559280515,
-0.2598329186,
-0.1850276291,
0.426317066,
-0.5619577765,
-0.2658594549,
-0.7943931818,
0.3577319384,
-0.4967299104,
0.7629107237,
-0.3709737957,
0.318653971,
-0.1442091465,
-... |
http://arxiv.org/abs/2406.16374v1 | http://arxiv.org/pdf/2406.16374v1 | KEHRL: Learning Knowledge-Enhanced Language Representations with Hierarchical Reinforcement Learning | 2024-06-24T07:32:35 | 2024-06-24T07:32:35 | Dongyang Li, Taolin Zhang, Longtao Huang, Chengyu Wang, Xiaofeng He, Hui Xue | cs.CL | cs.CL | Knowledge-enhanced pre-trained language models (KEPLMs) leverage relation
triples from knowledge graphs (KGs) and integrate these external data sources
into language models via self-supervised learning. Previous works treat
knowledge enhancement as two independent operations, i.e., knowledge injection
and knowledge int... | [
-0.1824518442,
-0.9643034339,
0.6749805212,
0.0665427223,
-0.0129826115,
0.4394758642,
-0.2196204364,
-0.5103008151,
-0.2005347908,
0.5418524742,
-0.1259690076,
-0.6292107105,
-0.6581119299,
-0.179287836,
0.1946656108,
0.9928064942,
-0.0224090889,
-0.0199395269,
-0.2516896129,
... |
http://arxiv.org/abs/2406.16372v1 | http://arxiv.org/pdf/2406.16372v1 | UniPSDA: Unsupervised Pseudo Semantic Data Augmentation for Zero-Shot Cross-Lingual Natural Language Understanding | 2024-06-24T07:27:01 | 2024-06-24T07:27:01 | Dongyang Li, Taolin Zhang, Jiali Deng, Longtao Huang, Chengyu Wang, Xiaofeng He, Hui Xue | cs.CL | cs.CL | Cross-lingual representation learning transfers knowledge from resource-rich
data to resource-scarce ones to improve the semantic understanding abilities of
different languages. However, previous works rely on shallow unsupervised data
generated by token surface matching, regardless of the global context-aware
semantic... | [
-0.2870753407,
-0.5330563784,
0.4687471092,
0.0500326827,
-0.8430768251,
0.1955859363,
-0.1066920012,
-0.3886804581,
0.413357228,
0.4957407415,
-0.5946727395,
-0.4845659137,
-0.2891481221,
0.0690363422,
-0.0891506448,
0.8253082633,
-0.0109769674,
0.3908719122,
-0.0519962832,
-0... |
http://arxiv.org/abs/2406.16356v1 | http://arxiv.org/pdf/2406.16356v1 | Evaluation of Instruction-Following Ability for Large Language Models on Story-Ending Generation | 2024-06-24T06:53:36 | 2024-06-24T06:53:36 | Rem Hida, Junki Ohmura, Toshiyuki Sekiya | cs.CL | cs.CL | Instruction-tuned Large Language Models (LLMs) have achieved remarkable
performance across various benchmark tasks. While providing instructions to
LLMs for guiding their generations is user-friendly, assessing their
instruction-following capabilities is still unclarified due to a lack of
evaluation metrics. In this pa... | [
-0.3300471902,
-0.467626214,
0.6456438899,
0.1592724919,
-0.2907617092,
0.3967039585,
-0.4286509156,
-0.218798086,
-0.2980012298,
0.5561640263,
-0.96280092,
-0.6893981099,
-0.6226198673,
0.2162717432,
-0.3406653702,
1.0325013399,
-0.2717185915,
-0.1225196645,
-0.4305339158,
0.1... |
http://arxiv.org/abs/2406.16342v1 | http://arxiv.org/pdf/2406.16342v1 | ADVSCORE: A Metric for the Evaluation and Creation of Adversarial Benchmarks | 2024-06-24T06:27:47 | 2024-06-24T06:27:47 | Yoo Yeon Sung, Eve Fleisig, Ishani Mondal, Jordan Lee Boyd-Graber | cs.CL | cs.CL | Adversarial benchmarks validate model abilities by providing samples that
fool models but not humans. However, despite the proliferation of datasets that
claim to be adversarial, there does not exist an established metric to evaluate
how adversarial these datasets are. To address this lacuna, we introduce
ADVSCORE, a m... | [
-0.364315182,
-0.8114119768,
0.2796439826,
-0.418964237,
0.2351774722,
0.2562462389,
0.0268205516,
0.1369445175,
-0.3345234096,
0.4745818675,
-0.4139665663,
-0.5311849713,
-0.5204643011,
-0.4190402329,
-0.6628235579,
1.2830832005,
0.0473606288,
0.0293837469,
-0.5324320197,
-0.2... |
http://arxiv.org/abs/2406.16341v1 | http://arxiv.org/pdf/2406.16341v1 | EHRCon: Dataset for Checking Consistency between Unstructured Notes and Structured Tables in Electronic Health Records | 2024-06-24T06:26:50 | 2024-06-24T06:26:50 | Yeonsu Kwon, Jiho Kim, Gyubok Lee, Seongsu Bae, Daeun Kyung, Wonchul Cha, Tom Pollard, Alistair Johnson, Edward Choi | cs.CL | cs.CL | Electronic Health Records (EHRs) are integral for storing comprehensive
patient medical records, combining structured data (e.g., medications) with
detailed clinical notes (e.g., physician notes). These elements are essential
for straightforward data retrieval and provide deep, contextual insights into
patient care. Ho... | [
0.2752124071,
-0.8907455802,
0.7646402717,
0.0799734369,
-0.2781180441,
-0.0733938962,
-0.1038343161,
-0.5252315998,
0.3639970422,
0.7672125101,
-0.0737050176,
-0.8976604342,
-0.3767599165,
0.254586935,
-0.3009307086,
1.4290369749,
-0.4868289828,
0.0163958389,
-0.1256680787,
-0... |
http://arxiv.org/abs/2406.16332v1 | http://arxiv.org/pdf/2406.16332v1 | DemoRank: Selecting Effective Demonstrations for Large Language Models in Ranking Task | 2024-06-24T06:10:13 | 2024-06-24T06:10:13 | Wenhan Liu, Yutao Zhu, Zhicheng Dou | cs.IR | cs.IR, cs.CL | Recently, there has been increasing interest in applying large language
models (LLMs) as zero-shot passage rankers. However, few studies have explored
how to select appropriate in-context demonstrations for the passage ranking
task, which is the focus of this paper. Previous studies mainly apply a
demonstration retriev... | [
-0.1953869015,
-0.5248978138,
0.6011096239,
0.272536993,
-0.2463044971,
-0.0104615996,
-0.1895094365,
-0.1228280291,
0.0530547574,
0.2415626198,
-0.2035916895,
-0.7644815445,
-0.7260627151,
-0.0286514591,
-0.4885528684,
1.0838329792,
0.1977099925,
0.0545719936,
-0.5844348073,
-... |
http://arxiv.org/abs/2406.16330v1 | http://arxiv.org/pdf/2406.16330v1 | Pruning via Merging: Compressing LLMs via Manifold Alignment Based Layer Merging | 2024-06-24T05:57:55 | 2024-06-24T05:57:55 | Deyuan Liu, Zhanyue Qin, Hairu Wang, Zhao Yang, Zecheng Wang, Fangying Rong, Qingbin Liu, Yanchao Hao, Xi Chen, Cunhang Fan, Zhao Lv, Zhiying Tu, Dianhui Chu, Bo Li, Dianbo Sui | cs.CL | cs.CL, cs.AI | While large language models (LLMs) excel in many domains, their complexity
and scale challenge deployment in resource-limited environments. Current
compression techniques, such as parameter pruning, often fail to effectively
utilize the knowledge from pruned parameters. To address these challenges, we
propose Manifold-... | [
-0.4355299473,
-0.4598167241,
0.4651080668,
-0.0836342648,
-0.5420277715,
0.1686717868,
-0.1909952462,
-0.1389096379,
0.0853011161,
0.5039885044,
-0.4713256061,
-0.4744227529,
-0.5047113299,
-0.3121652305,
-0.1472039372,
0.5937985778,
0.037937969,
0.0050970642,
-0.5445895791,
-... |
http://arxiv.org/abs/2406.16320v1 | http://arxiv.org/pdf/2406.16320v1 | What Do VLMs NOTICE? A Mechanistic Interpretability Pipeline for Noise-free Text-Image Corruption and Evaluation | 2024-06-24T05:13:19 | 2024-06-24T05:13:19 | Michal Golovanevsky, William Rudman, Vedant Palit, Ritambhara Singh, Carsten Eickhoff | cs.CL | cs.CL | Vision-Language Models (VLMs) have gained community-spanning prominence due
to their ability to integrate visual and textual inputs to perform complex
tasks. Despite their success, the internal decision-making processes of these
models remain opaque, posing challenges in high-stakes applications. To address
this, we in... | [
-0.0363377817,
-0.5439431071,
0.7569574714,
0.582945168,
-0.5360784531,
-0.3793141544,
0.1543470323,
-0.7036703229,
0.134634763,
0.7409651279,
-0.8433693647,
-0.2216978967,
-0.7795146108,
-0.0896226168,
-0.6801543832,
1.0516190529,
0.0686821938,
-0.0327484459,
-0.4558541477,
-0... |
http://arxiv.org/abs/2406.16319v1 | http://arxiv.org/pdf/2406.16319v1 | Modelled Multivariate Overlap: A method for measuring vowel merger | 2024-06-24T04:56:26 | 2024-06-24T04:56:26 | Irene Smith, Morgan Sonderegger, The Spade Consortium | cs.CL | cs.CL | This paper introduces a novel method for quantifying vowel overlap. There is
a tension in previous work between using multivariate measures, such as those
derived from empirical distributions, and the ability to control for unbalanced
data and extraneous factors, as is possible when using fitted model parameters.
The m... | [
-0.5763385892,
-0.6648150086,
0.5652653575,
0.4238680601,
0.142277658,
0.2857616842,
-0.1295336783,
-0.807126224,
0.4685949981,
0.0808430389,
-0.3117388487,
-0.3894594908,
-0.7045078278,
-0.2157280743,
-0.3310218453,
0.0485976674,
-0.1130864248,
0.1596471667,
0.0513456799,
-0.1... |
http://arxiv.org/abs/2406.16316v1 | http://arxiv.org/pdf/2406.16316v1 | Does Cross-Cultural Alignment Change the Commonsense Morality of Language Models? | 2024-06-24T04:50:12 | 2024-06-24T04:50:12 | Yuu Jinnai | cs.CL | cs.CL, cs.AI, cs.CY, cs.LG | Alignment of the language model with human preferences is a common approach
to making a language model useful to end users. However, most alignment work is
done in English, and human preference datasets are dominated by English,
reflecting only the preferences of English-speaking annotators. Nevertheless,
it is common ... | [
-0.2409341186,
-0.3508534431,
0.4285288155,
-0.1666540802,
-0.3913509548,
-0.3197911978,
-0.37999174,
-0.6595631838,
0.0740597099,
0.7618665099,
-0.3638179004,
-0.1879558116,
-0.4894379377,
0.2539782524,
0.2959988713,
1.1297930479,
-0.0155317625,
0.212785244,
-0.4513012767,
-0.... |
http://arxiv.org/abs/2406.16315v1 | http://arxiv.org/pdf/2406.16315v1 | Song Data Cleansing for End-to-End Neural Singer Diarization Using Neural Analysis and Synthesis Framework | 2024-06-24T04:48:29 | 2024-06-24T04:48:29 | Hokuto Munakata, Ryo Terashima, Yusuke Fujita | eess.AS | eess.AS, cs.CL, cs.SD | We propose a data cleansing method that utilizes a neural analysis and
synthesis (NANSY++) framework to train an end-to-end neural diarization model
(EEND) for singer diarization. Our proposed model converts song data with
choral singing which is commonly contained in popular music and unsuitable for
generating a simul... | [
-0.6566883922,
-0.5923986435,
-0.0541760959,
0.5049499273,
-0.1797025949,
0.1492871642,
-0.2934377193,
-0.7148723602,
0.7178384066,
0.5785831213,
-0.8184175491,
-0.5078821778,
-0.3141046762,
0.1616568416,
-0.8270209432,
1.0670309067,
0.3810817003,
0.2960508466,
-0.0846450254,
0... |
http://arxiv.org/abs/2406.16308v1 | http://arxiv.org/pdf/2406.16308v1 | Anomaly Detection of Tabular Data Using LLMs | 2024-06-24T04:17:03 | 2024-06-24T04:17:03 | Aodong Li, Yunhan Zhao, Chen Qiu, Marius Kloft, Padhraic Smyth, Maja Rudolph, Stephan Mandt | cs.LG | cs.LG, cs.AI, cs.CL | Large language models (LLMs) have shown their potential in long-context
understanding and mathematical reasoning. In this paper, we study the problem
of using LLMs to detect tabular anomalies and show that pre-trained LLMs are
zero-shot batch-level anomaly detectors. That is, without extra
distribution-specific model f... | [
-0.2224977463,
-1.1738128662,
0.8184396029,
-0.1745254099,
-0.2023123801,
0.0349613093,
-0.2475197166,
-0.1190990806,
0.0326714031,
0.6341068745,
-0.7713939548,
-0.9104192853,
-0.4469832182,
0.057637684,
-0.6689965725,
0.8920606375,
-0.247156173,
-0.1896685362,
-0.1851586998,
0... |
http://arxiv.org/abs/2406.16306v1 | http://arxiv.org/pdf/2406.16306v1 | Cascade Reward Sampling for Efficient Decoding-Time Alignment | 2024-06-24T04:08:35 | 2024-06-24T04:08:35 | Bolian Li, Yifan Wang, Ananth Grama, Ruqi Zhang | cs.CL | cs.CL, cs.LG, stat.ML | Aligning large language models (LLMs) with human preferences is critical for
their deployment. Recently, decoding-time alignment has emerged as an effective
plug-and-play technique that requires no fine-tuning of model parameters.
However, generating text that achieves both high reward and high likelihood
remains a sig... | [
-0.1479103267,
-0.1811329722,
0.445522964,
0.3887684941,
-0.3800342679,
0.1283708364,
-0.2958082259,
-0.2746101618,
0.3581516147,
0.2272560298,
-0.7560261488,
-0.4103433192,
-0.5534315109,
0.0207645968,
-0.1859519929,
1.2266132832,
0.0563537031,
-0.2665576637,
-0.1092521846,
0.... |
http://arxiv.org/abs/2406.17808v1 | http://arxiv.org/pdf/2406.17808v1 | Training-Free Exponential Extension of Sliding Window Context with Cascading KV Cache | 2024-06-24T03:59:17 | 2024-06-24T03:59:17 | Jeffrey Willette, Heejun Lee, Youngwan Lee, Myeongjae Jeon, Sung Ju Hwang | cs.CL | cs.CL, cs.AI, cs.LG | The context window within a transformer provides a form of active memory for
the current task, which can be useful for few-shot learning and conditional
generation, both which depend heavily on previous context tokens. However, as
the context length grows, the computational cost increases quadratically.
Recent works ha... | [
-0.2232467234,
-0.5185567737,
0.3619553149,
0.3476544917,
-0.4618285298,
-0.0072114496,
-0.5230640769,
-0.3389131725,
0.3359222412,
0.2433634996,
-0.6142354012,
-0.2886428833,
-0.3689971864,
-0.3902769387,
-0.3755940497,
0.9606518745,
-0.2873640358,
-0.1584108174,
-0.1590373069,
... |
http://arxiv.org/abs/2406.16299v1 | http://arxiv.org/pdf/2406.16299v1 | Compensate Quantization Errors: Make Weights Hierarchical to Compensate Each Other | 2024-06-24T03:52:52 | 2024-06-24T03:52:52 | Yifei Gao, Jie Ou, Lei Wang, Yuting Xiao, Zhiyuan Xiang, Ruiting Dai, Jun Cheng | cs.CL | cs.CL, cs.AI, F.2.3 | Emergent Large Language Models (LLMs) use their extraordinary performance and
powerful deduction capacity to discern from traditional language models.
However, the expenses of computational resources and storage for these LLMs are
stunning, quantization then arises as a trending conversation. To address
accuracy decay ... | [
0.3192875385,
-0.4660958052,
0.462109983,
0.2178084999,
-0.1672282368,
0.3347397447,
-0.2339897007,
-0.2223713696,
0.0321617872,
0.2581453323,
-0.4254512489,
-0.1721710265,
-0.5533286929,
-0.0933368802,
-0.1105143949,
0.9913187623,
0.0756719038,
0.2252160013,
-0.3217163682,
-0.... |
http://arxiv.org/abs/2406.16294v1 | http://arxiv.org/pdf/2406.16294v1 | LangSuitE: Planning, Controlling and Interacting with Large Language Models in Embodied Text Environments | 2024-06-24T03:36:29 | 2024-06-24T03:36:29 | Zixia Jia, Mengmeng Wang, Baichen Tong, Song-Chun Zhu, Zilong Zheng | cs.CL | cs.CL, cs.AI | Recent advances in Large Language Models (LLMs) have shown inspiring
achievements in constructing autonomous agents that rely on language
descriptions as inputs. However, it remains unclear how well LLMs can function
as few-shot or zero-shot embodied agents in dynamic interactive environments.
To address this gap, we i... | [
-0.4155195355,
-0.8037637472,
0.7767887115,
0.2258398235,
-0.0886941701,
0.0111636212,
-0.1786703318,
-0.4350143671,
0.400834173,
0.6708111763,
-0.6462396383,
-0.6179250479,
-0.3902308345,
-0.0821410269,
-0.4516632855,
0.5584087372,
-0.4949731529,
-0.5249297023,
-0.0721015856,
... |
http://arxiv.org/abs/2406.16293v1 | http://arxiv.org/pdf/2406.16293v1 | Combining Supervised Learning and Reinforcement Learning for Multi-Label Classification Tasks with Partial Labels | 2024-06-24T03:36:19 | 2024-06-24T03:36:19 | Zixia Jia, Junpeng Li, Shichuan Zhang, Anji Liu, Zilong Zheng | cs.CL | cs.CL, cs.AI | Traditional supervised learning heavily relies on human-annotated datasets,
especially in data-hungry neural approaches. However, various tasks, especially
multi-label tasks like document-level relation extraction, pose challenges in
fully manual annotation due to the specific domain knowledge and large class
sets. The... | [
-0.5976780057,
-0.7406007648,
0.2927677333,
0.3657637239,
-0.0342748091,
0.3395284116,
-0.1503334194,
-0.4810968041,
0.111026518,
0.6011670232,
-0.6304263473,
-0.4374983013,
-0.8115853071,
0.0320335105,
-0.0003028469,
1.4883308411,
-0.1254073381,
-0.230208233,
-0.4456976056,
-0... |
http://arxiv.org/abs/2406.16288v1 | http://arxiv.org/pdf/2406.16288v1 | PlagBench: Exploring the Duality of Large Language Models in Plagiarism Generation and Detection | 2024-06-24T03:29:53 | 2024-06-24T03:29:53 | Jooyoung Lee, Toshini Agrawal, Adaku Uchendu, Thai Le, Jinghui Chen, Dongwon Lee | cs.CL | cs.CL | Recent literature has highlighted potential risks to academic integrity
associated with large language models (LLMs), as they can memorize parts of
training instances and reproduce them in the generated texts without proper
attribution. In addition, given their capabilities in generating high-quality
texts, plagiarists... | [
-0.1786803454,
-0.842453897,
0.6789442301,
-0.0088942172,
-0.5207303166,
0.2395083755,
-0.091793932,
-0.2354843765,
-0.1024402156,
0.770306766,
-0.2377755493,
-0.514567256,
-0.5500674844,
0.1456744373,
-0.7711322904,
1.239439249,
-0.2016512603,
0.0450557433,
-0.4983015656,
-0.1... |
http://arxiv.org/abs/2406.16275v1 | http://arxiv.org/pdf/2406.16275v1 | Investigating the Influence of Prompt-Specific Shortcuts in AI Generated Text Detection | 2024-06-24T02:50:09 | 2024-06-24T02:50:09 | Choonghyun Park, Hyuhng Joon Kim, Junyeob Kim, Youna Kim, Taeuk Kim, Hyunsoo Cho, Hwiyeol Jo, Sang-goo Lee, Kang Min Yoo | cs.CL | cs.CL | AI Generated Text (AIGT) detectors are developed with texts from humans and
LLMs of common tasks. Despite the diversity of plausible prompt choices, these
datasets are generally constructed with a limited number of prompts. The lack
of prompt variation can introduce prompt-specific shortcut features that exist
in data ... | [
-0.4292580187,
-0.8029420376,
0.48049137,
-0.4134400189,
-0.2323245108,
0.0376661606,
-0.1778793782,
0.0379986875,
-0.1168533713,
0.3012463152,
-0.7873589396,
-0.5457085371,
-0.7631120682,
0.3047465086,
-0.2293763608,
1.2438173294,
0.064317517,
-0.2819658816,
0.0093757091,
0.02... |
http://arxiv.org/abs/2406.16264v1 | http://arxiv.org/pdf/2406.16264v1 | One Thousand and One Pairs: A "novel" challenge for long-context language models | 2024-06-24T02:03:57 | 2024-06-24T02:03:57 | Marzena Karpinska, Katherine Thai, Kyle Lo, Tanya Goyal, Mohit Iyyer | cs.CL | cs.CL, cs.AI | Synthetic long-context LLM benchmarks (e.g., "needle-in-the-haystack") test
only surface-level retrieval capabilities, but how well can long-context LLMs
retrieve, synthesize, and reason over information across book-length inputs? We
address this question by creating NoCha, a dataset of 1,001 minimally different
pairs ... | [
-0.3120425642,
-0.7178190947,
0.8299030662,
0.1031203866,
-0.113491267,
-0.0394295231,
-0.114018999,
-0.746396482,
-0.1574805379,
0.6975377798,
-0.5509295464,
-0.3791299164,
-0.2978722453,
0.4488465786,
-0.5140109062,
0.8062738776,
-0.1379374713,
-0.1819743067,
-0.6756469607,
-... |
http://arxiv.org/abs/2406.16254v1 | http://arxiv.org/pdf/2406.16254v1 | Confidence Regulation Neurons in Language Models | 2024-06-24T01:31:03 | 2024-06-24T01:31:03 | Alessandro Stolfo, Ben Wu, Wes Gurnee, Yonatan Belinkov, Xingyi Song, Mrinmaya Sachan, Neel Nanda | cs.LG | cs.LG, cs.AI, cs.CL | Despite their widespread use, the mechanisms by which large language models
(LLMs) represent and regulate uncertainty in next-token predictions remain
largely unexplored. This study investigates two critical components believed to
influence this uncertainty: the recently discovered entropy neurons and a new
set of comp... | [
-0.0938292295,
-0.4127804637,
0.2895623147,
0.2377757579,
0.0105496794,
-0.1254217774,
-0.3094022274,
-0.4510619342,
0.1787733436,
0.2356640995,
-0.2073103338,
-0.354326427,
-0.6303717494,
-0.0390441194,
-0.4709531665,
0.8448355794,
0.2850355208,
0.062815249,
-0.1170932725,
0.3... |
http://arxiv.org/abs/2406.16253v2 | http://arxiv.org/pdf/2406.16253v2 | LLMs Assist NLP Researchers: Critique Paper (Meta-)Reviewing | 2024-06-24T01:30:22 | 2024-06-25T18:04:38 | Jiangshu Du, Yibo Wang, Wenting Zhao, Zhongfen Deng, Shuaiqi Liu, Renze Lou, Henry Peng Zou, Pranav Narayanan Venkit, Nan Zhang, Mukund Srinath, Haoran Ranran Zhang, Vipul Gupta, Yinghui Li, Tao Li, Fei Wang, Qin Liu, Tianlin Liu, Pengzhi Gao, Congying Xia, Chen Xing, Jiayang Cheng, Zhaowei Wang, Ying Su, Raj Sanjay Sh... | cs.CL | cs.CL | This work is motivated by two key trends. On one hand, large language models
(LLMs) have shown remarkable versatility in various generative tasks such as
writing, drawing, and question answering, significantly reducing the time
required for many routine tasks. On the other hand, researchers, whose work is
not only time... | [
-0.4538398385,
-0.4555823803,
0.6671296358,
0.3120366335,
-0.3012581468,
0.1907170862,
-0.2194630951,
-0.7169146538,
0.4739282131,
0.5777474642,
-0.4896576703,
-0.4586255252,
-0.5505555272,
0.7055103779,
-0.432883352,
1.2991788387,
-0.1847742498,
0.2087302953,
-0.3832235932,
0.... |
http://arxiv.org/abs/2406.16985v1 | http://arxiv.org/pdf/2406.16985v1 | Unveiling LLM Mechanisms Through Neural ODEs and Control Theory | 2024-06-23T22:56:34 | 2024-06-23T22:56:34 | Yukun Zhang | cs.LG | cs.LG, cs.AI, cs.CL | This study presents a novel approach that leverages Neural Ordinary
Differential Equations (Neural ODEs) to unravel the intricate relationships
between inputs and outputs in Large Language Models (LLMs), and employs robust
control to fine-tune outputs to meet predefined standards. Central to our
methodology is the tran... | [
-0.1951549649,
-0.8730137348,
0.611225307,
0.0567976385,
-0.0866619423,
0.0694115311,
-0.4096227586,
-0.2863612473,
-0.0655591264,
0.2248121202,
-0.5676303506,
-0.3385185003,
-0.8180468678,
-0.1027848944,
-0.4481863976,
0.924828589,
-0.327952683,
0.1438495517,
-0.2525386214,
0.... |
http://arxiv.org/abs/2406.16235v1 | http://arxiv.org/pdf/2406.16235v1 | Preference Tuning For Toxicity Mitigation Generalizes Across Languages | 2024-06-23T22:53:47 | 2024-06-23T22:53:47 | Xiaochen Li, Zheng-Xin Yong, Stephen H. Bach | cs.CL | cs.CL, cs.AI, cs.CR, cs.LG | Detoxifying multilingual Large Language Models (LLMs) has become crucial due
to their increasing global use. In this work, we explore zero-shot
cross-lingual generalization of preference tuning in detoxifying LLMs. Unlike
previous studies that show limited cross-lingual generalization for other
safety tasks, we demonst... | [
-0.0634682178,
-0.3964953125,
0.534557879,
0.2930718958,
-0.1790050715,
-0.0823175684,
-0.3479352593,
-0.3244237304,
-0.0139646353,
0.595051229,
-0.5028053522,
-0.5949738026,
-0.4811839759,
0.2947075665,
0.4996656477,
1.1823785305,
-0.0236413721,
0.3081949055,
0.2393551022,
-0.... |
http://arxiv.org/abs/2406.16229v1 | http://arxiv.org/pdf/2406.16229v1 | Multi-Objective Linguistic Control of Large Language Models | 2024-06-23T21:56:48 | 2024-06-23T21:56:48 | Dang Nguyen, Jiuhai Chen, Tianyi Zhou | cs.CL | cs.CL | Large language models (LLMs), despite their breakthroughs on many challenging
benchmark tasks, lean to generate verbose responses and lack the
controllability of output complexity, which is usually preferred by human users
in practice. In this paper, we study how to precisely control multiple
linguistic complexities of... | [
-0.7457177639,
-1.0325043201,
0.314901799,
0.108167395,
-0.0770695359,
-0.0097482773,
-0.7572936416,
-0.1582874656,
-0.1220166385,
0.6745579839,
-0.5962856412,
-0.475633949,
-0.8411391973,
0.1251746416,
-0.0211878326,
1.067735672,
-0.0866200477,
-0.1013056114,
0.3052960336,
-0.... |
http://arxiv.org/abs/2406.16223v1 | http://arxiv.org/pdf/2406.16223v1 | Continuous Output Personality Detection Models via Mixed Strategy Training | 2024-06-23T21:32:15 | 2024-06-23T21:32:15 | Rong Wang, Kun Sun | cs.CL | cs.CL, cs.AI | The traditional personality models only yield binary results. This paper
presents a novel approach for training personality detection models that
produce continuous output values, using mixed strategies. By leveraging the
PANDORA dataset, which includes extensive personality labeling of Reddit
comments, we developed mo... | [
-0.6068514585,
-0.5976865292,
0.8622221351,
0.3013269603,
-0.216482535,
0.098078467,
-0.5303183794,
-0.2346518636,
0.252579391,
0.691218257,
-1.0491224527,
-0.4882166684,
-0.7847858667,
-0.0858234614,
-0.2918175757,
1.2798193693,
0.4720756114,
0.221961543,
0.0379519686,
-0.2630... |
http://arxiv.org/abs/2406.16203v3 | http://arxiv.org/pdf/2406.16203v3 | LLMs' Classification Performance is Overclaimed | 2024-06-23T19:49:10 | 2024-07-03T13:18:50 | Hanzi Xu, Renze Lou, Jiangshu Du, Vahid Mahzoon, Elmira Talebianaraki, Zhuoan Zhou, Elizabeth Garrison, Slobodan Vucetic, Wenpeng Yin | cs.CL | cs.CL | In many classification tasks designed for AI or human to solve, gold labels
are typically included within the label space by default, often posed as "which
of the following is correct?" This standard setup has traditionally highlighted
the strong performance of advanced AI, particularly top-performing Large
Language Mo... | [
-0.388644129,
-0.5332320333,
0.5961954594,
-0.040927995,
0.0672712699,
0.1886244416,
-0.4189316332,
-0.6982915401,
-0.0057845665,
0.5358778834,
-0.4882481694,
-0.6319286823,
-0.8124538064,
0.3052126169,
-0.1773294508,
1.1739510298,
-0.1025365293,
-0.3201270103,
-0.6157220602,
-... |
http://arxiv.org/abs/2406.16201v1 | http://arxiv.org/pdf/2406.16201v1 | Blind Baselines Beat Membership Inference Attacks for Foundation Models | 2024-06-23T19:40:11 | 2024-06-23T19:40:11 | Debeshee Das, Jie Zhang, Florian Tramèr | cs.CR | cs.CR, cs.CL, cs.LG | Membership inference (MI) attacks try to determine if a data sample was used
to train a machine learning model. For foundation models trained on unknown Web
data, MI attacks can be used to detect copyrighted training materials, measure
test set contamination, or audit machine unlearning. Unfortunately, we find
that eva... | [
-0.4280219078,
-0.4816733599,
0.0786075071,
-0.0811541304,
0.0219103824,
0.0553311147,
0.5696727633,
-0.6384062767,
0.1142005101,
0.6348827481,
-0.891065836,
-0.2582103908,
-0.6084757447,
0.2719309032,
-0.6805739403,
0.7829071879,
0.3021834195,
0.3331297636,
-0.3892273307,
-0.4... |
http://arxiv.org/abs/2406.16176v1 | http://arxiv.org/pdf/2406.16176v1 | GraphEval2000: Benchmarking and Improving Large Language Models on Graph Datasets | 2024-06-23T18:01:56 | 2024-06-23T18:01:56 | Qiming Wu, Zichen Chen, Will Corcoran, Misha Sra, Ambuj K. Singh | cs.AI | cs.AI, cs.CL, cs.LG, H.2.8, I.2.6, I.2.7 | Large language models (LLMs) have achieved remarkable success in natural
language processing (NLP), demonstrating significant capabilities in processing
and understanding text data. However, recent studies have identified
limitations in LLMs' ability to reason about graph-structured data. To address
this gap, we introd... | [
-0.1040917188,
-0.9363003373,
0.464296788,
0.2507961392,
-0.0274725016,
0.5666102171,
-0.2316184938,
-0.3717192709,
-0.2215523869,
0.4717330635,
-0.4388462007,
-1.0588662624,
-0.626967907,
0.1988140196,
-0.3588575125,
0.6984342337,
-0.2571189702,
0.1321841776,
-0.434604764,
-0.... |
http://arxiv.org/abs/2406.16167v1 | http://arxiv.org/pdf/2406.16167v1 | FS-RAG: A Frame Semantics Based Approach for Improved Factual Accuracy in Large Language Models | 2024-06-23T17:18:19 | 2024-06-23T17:18:19 | Harish Tayyar Madabushi | cs.CL | cs.CL | We present a novel extension to Retrieval Augmented Generation with the goal
of mitigating factual inaccuracies in the output of large language models.
Specifically, our method draws on the cognitive linguistic theory of frame
semantics for the indexing and retrieval of factual information relevant to
helping large lan... | [
-0.3440160155,
-1.0055315495,
0.7782665491,
-0.1542002857,
-0.0522868931,
0.1718270034,
-0.3124721944,
-0.3883502185,
0.4821092486,
0.5954364538,
-0.3723073006,
-0.2151243985,
-0.5598470569,
-0.0361804515,
-0.6758399606,
1.0000523329,
0.0803249478,
-0.0824974254,
-0.2271659076,
... |
http://arxiv.org/abs/2406.16152v1 | http://arxiv.org/pdf/2406.16152v1 | Towards Region-aware Bias Evaluation Metrics | 2024-06-23T16:26:27 | 2024-06-23T16:26:27 | Angana Borah, Aparna Garimella, Rada Mihalcea | cs.CL | cs.CL | When exposed to human-generated data, language models are known to learn and
amplify societal biases. While previous works introduced benchmarks that can be
used to assess the bias in these models, they rely on assumptions that may not
be universally true. For instance, a gender bias dimension commonly used by
these me... | [
-0.6174015999,
-0.4281633496,
0.4507652521,
0.1372405142,
0.0611427501,
0.3198189437,
-0.2642600834,
-0.0260407552,
0.1436077058,
0.517137289,
-0.7988214493,
-0.5190286636,
-0.7564113736,
-0.2216198891,
-0.1105679348,
0.8431082964,
0.0620851777,
0.2874402702,
-0.2227546722,
-0.... |
http://arxiv.org/abs/2407.02518v1 | http://arxiv.org/pdf/2407.02518v1 | INDICT: Code Generation with Internal Dialogues of Critiques for Both Security and Helpfulness | 2024-06-23T15:55:07 | 2024-06-23T15:55:07 | Hung Le, Yingbo Zhou, Caiming Xiong, Silvio Savarese, Doyen Sahoo | cs.SE | cs.SE, cs.AI, cs.CL, cs.CR, cs.MA, cs.PL | Large language models (LLMs) for code are typically trained to align with
natural language instructions to closely follow their intentions and
requirements. However, in many practical scenarios, it becomes increasingly
challenging for these models to navigate the intricate boundary between
helpfulness and safety, espec... | [
-0.1736572981,
-0.2942935228,
0.286929518,
0.2436419874,
-0.3600537777,
0.2993791103,
-0.3890497684,
-0.4515932798,
-0.0415793061,
0.5602582097,
-0.4683433175,
-0.5612429976,
-0.3575201631,
0.1873360276,
-0.1316285878,
1.525164485,
0.0679392964,
0.4064707756,
-0.1020758152,
0.0... |
http://arxiv.org/abs/2406.16144v1 | http://arxiv.org/pdf/2406.16144v1 | Chain-of-Probe: Examing the Necessity and Accuracy of CoT Step-by-Step | 2024-06-23T15:50:22 | 2024-06-23T15:50:22 | Zezhong Wang, Xingshan Zeng, Weiwen Liu, Yufei Wang, Liangyou Li, Yasheng Wang, Lifeng Shang, Xin Jiang, Qun Liu, Kam-Fai Wong | cs.CL | cs.CL | Current research found the issue of Early Answering in large language models
(LLMs), where the models already have an answer before generating the
Chain-of-Thought (CoT). This phenomenon suggests a potential lack of necessary
dependency between the predicted answer and the reasoning process.
Consequently, two important... | [
-0.6492057443,
-0.9345893264,
1.0567370653,
-0.063845627,
-0.1739519835,
0.3180236816,
0.0036727206,
-0.255455792,
0.1077528968,
0.5821540952,
-0.4311719835,
-0.1638981849,
-0.6184047461,
0.020273352,
-0.5124238133,
0.7754813433,
-0.2819733024,
0.1214608625,
-0.1750403196,
-0.0... |
http://arxiv.org/abs/2406.16135v1 | http://arxiv.org/pdf/2406.16135v1 | Crosslingual Capabilities and Knowledge Barriers in Multilingual Large Language Models | 2024-06-23T15:15:17 | 2024-06-23T15:15:17 | Lynn Chua, Badih Ghazi, Yangsibo Huang, Pritish Kamath, Ravi Kumar, Pasin Manurangsi, Amer Sinha, Chulin Xie, Chiyuan Zhang | cs.CL | cs.CL, cs.LG | Large language models (LLMs) are typically multilingual due to pretraining on
diverse multilingual corpora. But can these models relate corresponding
concepts across languages, effectively being crosslingual? This study evaluates
six state-of-the-art LLMs on inherently crosslingual tasks. We observe that
while these mo... | [
-0.3774853945,
-0.670492053,
0.5115563869,
0.3258931041,
0.1353872269,
0.2693805695,
-0.514900744,
-0.7499085069,
-0.2388876379,
0.3103857636,
-0.6146087646,
-0.3423441648,
-0.5924227834,
0.0879212692,
-0.0820903108,
0.6931049824,
-0.3862426877,
0.1640136093,
-0.655807972,
-0.2... |
http://arxiv.org/abs/2406.16120v1 | http://arxiv.org/pdf/2406.16120v1 | Contextualized End-to-end Automatic Speech Recognition with Intermediate Biasing Loss | 2024-06-23T14:22:59 | 2024-06-23T14:22:59 | Muhammad Shakeel, Yui Sudo, Yifan Peng, Shinji Watanabe | eess.AS | eess.AS, cs.CL, cs.SD | Contextualized end-to-end automatic speech recognition has been an active
research area, with recent efforts focusing on the implicit learning of
contextual phrases based on the final loss objective. However, these approaches
ignore the useful contextual knowledge encoded in the intermediate layers. We
hypothesize that... | [
-0.3193511963,
-0.5827121735,
0.2642605603,
0.0345511883,
-0.3823420405,
-0.2467010468,
-0.3922596276,
-0.2731100619,
0.0922040716,
0.2996368706,
-0.4851069748,
-0.2851668596,
-0.646397531,
0.0619658381,
-0.6256878972,
0.9041374326,
-0.1555554271,
0.3071739376,
-0.0832080916,
-... |
http://arxiv.org/abs/2406.16107v1 | http://arxiv.org/pdf/2406.16107v1 | Decoder-only Architecture for Streaming End-to-end Speech Recognition | 2024-06-23T13:50:08 | 2024-06-23T13:50:08 | Emiru Tsunoo, Hayato Futami, Yosuke Kashiwagi, Siddhant Arora, Shinji Watanabe | eess.AS | eess.AS, cs.CL | Decoder-only language models (LMs) have been successfully adopted for
speech-processing tasks including automatic speech recognition (ASR). The LMs
have ample expressiveness and perform efficiently. This efficiency is a
suitable characteristic for streaming applications of ASR. In this work, we
propose to use a decoder... | [
-0.3817594349,
-0.5074466467,
0.2001413554,
-0.0176636744,
-0.8289600015,
-0.0376640484,
-0.3788763285,
-0.3799573779,
0.5492908359,
0.4057944715,
-0.5131993294,
-0.1979326904,
-0.7267519236,
0.275370419,
-0.7845563293,
0.3781619072,
0.1047409624,
0.1025526822,
0.039524816,
0.0... |
http://arxiv.org/abs/2406.17807v1 | http://arxiv.org/pdf/2406.17807v1 | Enhancing Commentary Strategies for Imperfect Information Card Games: A Study of Large Language Models in Guandan Commentary | 2024-06-23T11:58:26 | 2024-06-23T11:58:26 | Meiling Tao. Xuechen Liang, Yiling Tao, Tianyu Shi | cs.CL | cs.CL, cs.AI | Recent advancements in large language models (LLMs) have unlocked the
potential for generating high-quality game commentary. However, producing
insightful and engaging commentary for complex games with incomplete
information remains a significant challenge. In this paper, we introduce a
novel commentary method that com... | [
-0.2323856801,
-0.6731436253,
0.4441152215,
0.5306965113,
-0.5536484718,
-0.0211087763,
-0.7674003243,
-0.4111979604,
0.6131142974,
0.411621958,
-0.6468573809,
-0.5045026541,
-0.192684859,
-0.109201923,
0.3515211344,
1.5148261786,
0.14784199,
0.3625198007,
-0.1535671949,
-0.069... |
http://arxiv.org/abs/2406.16086v1 | http://arxiv.org/pdf/2406.16086v1 | SEAM: A Stochastic Benchmark for Multi-Document Tasks | 2024-06-23T11:57:53 | 2024-06-23T11:57:53 | Gili Lior, Avi Caciularu, Arie Cattan, Shahar Levy, Ori Shapira, Gabriel Stanovsky | cs.CL | cs.CL | Various tasks, such as summarization, multi-hop question answering, or
coreference resolution, are naturally phrased over collections of real-world
documents. Such tasks present a unique set of challenges, revolving around the
lack of coherent narrative structure across documents, which often leads to
contradiction, om... | [
-0.1918358505,
-0.7823292613,
0.7132635117,
0.4237275124,
-0.1265220195,
0.271063745,
-0.5010378361,
-0.3112129271,
0.0657645017,
0.2194814384,
-0.771150887,
-0.574846983,
-1.009940505,
0.0144566922,
-0.379483819,
0.870402813,
-0.3498011827,
-0.0889652818,
-0.528785646,
-0.1540... |
http://arxiv.org/abs/2406.16079v1 | http://arxiv.org/pdf/2406.16079v1 | EERPD: Leveraging Emotion and Emotion Regulation for Improving Personality Detection | 2024-06-23T11:18:55 | 2024-06-23T11:18:55 | Zheng Li, Dawei Zhu, Qilong Ma, Weimin Xiong, Sujian Li | cs.CL | cs.CL, cs.AI | Personality is a fundamental construct in psychology, reflecting an
individual's behavior, thinking, and emotional patterns. Previous researches
have made some progress in personality detection, primarily by utilizing the
whole text to predict personality. However, these studies generally tend to
overlook psychological... | [
-0.7430445552,
-0.538254559,
0.5048217177,
0.3338696957,
-0.1298446506,
-0.0444351099,
-0.2649436593,
-0.6011750102,
0.2865784168,
0.6009231806,
-0.6173453331,
-0.3961759508,
-0.632347405,
0.2562069297,
-0.0745754912,
0.9304357171,
0.0357096829,
0.0174717996,
-0.0823970437,
0.0... |
http://arxiv.org/abs/2406.16078v1 | http://arxiv.org/pdf/2406.16078v1 | First Heuristic Then Rational: Dynamic Use of Heuristics in Language Model Reasoning | 2024-06-23T11:11:46 | 2024-06-23T11:11:46 | Yoichi Aoki, Keito Kudo, Tatsuki Kuribayashi, Shusaku Sone, Masaya Taniguchi, Keisuke Sakaguchi, Kentaro Inui | cs.CL | cs.CL | Multi-step reasoning is widely adopted in the community to explore the better
performance of language models (LMs). We report on the systematic strategy that
LMs use in this process. Our controlled experiments reveal that LMs rely more
heavily on heuristics, such as lexical overlap, in the earlier stages of
reasoning w... | [
-0.5010432005,
-0.7273699045,
0.9733358026,
0.1214404032,
0.0003684159,
0.5311040282,
-0.0744258463,
-0.4003339112,
-0.1739605367,
0.2779776752,
-0.8127667904,
-0.1730921119,
-0.5104818344,
0.021143632,
-0.4110182822,
0.3890686929,
-0.2662889659,
0.0883198231,
-0.0605843551,
-0... |
http://arxiv.org/abs/2406.16071v1 | http://arxiv.org/pdf/2406.16071v1 | Dancing in the syntax forest: fast, accurate and explainable sentiment analysis with SALSA | 2024-06-23T10:47:01 | 2024-06-23T10:47:01 | Carlos Gómez-Rodríguez, Muhammad Imran, David Vilares, Elena Solera, Olga Kellert | cs.CL | cs.CL, 68T50, I.2.7 | Sentiment analysis is a key technology for companies and institutions to
gauge public opinion on products, services or events. However, for large-scale
sentiment analysis to be accessible to entities with modest computational
resources, it needs to be performed in a resource-efficient way. While some
efficient sentimen... | [
-0.200218752,
-0.641934514,
0.3755320013,
0.8434936404,
-0.4809981287,
0.5182165504,
-0.381565094,
-0.3655894399,
0.6656466126,
0.2016386688,
-0.6523331404,
-0.4811279178,
-0.5109351277,
-0.1680587232,
-0.6509327292,
0.9701515436,
0.1963529438,
0.1216574684,
-0.1296172589,
-0.5... |
http://arxiv.org/abs/2406.16069v1 | http://arxiv.org/pdf/2406.16069v1 | FastMem: Fast Memorization of Prompt Improves Context Awareness of Large Language Models | 2024-06-23T10:36:35 | 2024-06-23T10:36:35 | Junyi Zhu, Shuochen Liu, Yu Yu, Bo Tang, Yibo Yan, Zhiyu Li, Feiyu Xiong, Tong Xu, Matthew B. Blaschko | cs.CL | cs.CL, cs.AI | Large language models (LLMs) excel in generating coherent text, but they
often struggle with context awareness, leading to inaccuracies in tasks
requiring faithful adherence to provided information. We introduce FastMem, a
novel method designed to enhance instruction fine-tuned LLMs' context awareness
through fast memo... | [
-0.1660734415,
-0.9459047914,
0.6444125772,
0.4652823508,
-0.2232191712,
-0.0520810075,
-0.4619481564,
-0.0149087477,
-0.1429252476,
0.4692956805,
-0.4345330894,
-0.6093334556,
-0.5173895955,
0.1408158988,
-0.155196175,
0.5873567462,
-0.0537643284,
0.0852195546,
-0.2853250802,
... |
http://arxiv.org/abs/2406.16061v1 | http://arxiv.org/pdf/2406.16061v1 | PORT: Preference Optimization on Reasoning Traces | 2024-06-23T09:51:06 | 2024-06-23T09:51:06 | Salem Lahlou, Abdalgader Abubaker, Hakim Hacid | cs.LG | cs.LG, cs.CL | Preference optimization methods have been successfully applied to improve not
only the alignment of large language models (LLMs) with human values, but also
specific natural language tasks such as summarization and stylistic
continuations. This paper proposes using preference optimization methods on
Chain-of-Thought st... | [
-0.6384996772,
-0.6321136355,
0.8208658099,
-0.0997919738,
-0.4901863039,
-0.0024510552,
-0.3178533912,
-0.2966713309,
0.3326242566,
0.877241075,
-0.6843619347,
-0.1604887247,
-0.2702816725,
0.1608034968,
0.024546694,
0.8917321563,
-0.1081804037,
0.007884671,
-0.1217520833,
-0.... |
http://arxiv.org/abs/2406.16033v1 | http://arxiv.org/pdf/2406.16033v1 | Unlocking the Future: Exploring Look-Ahead Planning Mechanistic Interpretability in Large Language Models | 2024-06-23T06:54:47 | 2024-06-23T06:54:47 | Tianyi Men, Pengfei Cao, Zhuoran Jin, Yubo Chen, Kang Liu, Jun Zhao | cs.CL | cs.CL | Planning, as the core module of agents, is crucial in various fields such as
embodied agents, web navigation, and tool using. With the development of large
language models (LLMs), some researchers treat large language models as
intelligent agents to stimulate and evaluate their planning capabilities.
However, the plann... | [
-0.3681189716,
-0.4654267728,
0.904622972,
0.1477841586,
-0.0553288572,
0.0052960669,
-0.1648387611,
-0.6324663758,
0.275146693,
0.6493213773,
-1.1840264797,
-0.2119691819,
-0.4424458742,
-0.0860867426,
-0.2708721161,
0.5541633368,
-0.2181113511,
-0.0365739167,
-0.0466050655,
0... |
http://arxiv.org/abs/2406.16030v1 | http://arxiv.org/pdf/2406.16030v1 | Zero-Shot Cross-Lingual NER Using Phonemic Representations for Low-Resource Languages | 2024-06-23T06:38:56 | 2024-06-23T06:38:56 | Jimin Sohn, Haeji Jung, Alex Cheng, Jooeon Kang, Yilin Du, David R. Mortensen | cs.CL | cs.CL, cs.AI | Existing zero-shot cross-lingual NER approaches require substantial prior
knowledge of the target language, which is impractical for low-resource
languages. In this paper, we propose a novel approach to NER using phonemic
representation based on the International Phonetic Alphabet (IPA) to bridge the
gap between repres... | [
-0.2851963937,
-0.2890929878,
0.257527858,
0.206538111,
-0.2209637761,
0.2492338717,
-0.2066542655,
-0.5320910811,
0.6556863785,
0.1849632561,
-0.1126994118,
-0.5804129243,
-0.3915624321,
0.3237521648,
0.0862149745,
0.3936239779,
0.2069457918,
0.1749211252,
0.3197077513,
-0.256... |
http://arxiv.org/abs/2406.16021v1 | http://arxiv.org/pdf/2406.16021v1 | Harvesting Events from Multiple Sources: Towards a Cross-Document Event Extraction Paradigm | 2024-06-23T06:01:11 | 2024-06-23T06:01:11 | Qiang Gao, Zixiang Meng, Bobo Li, Jun Zhou, Fei Li, Chong Teng, Donghong Ji | cs.CL | cs.CL, cs.AI | Document-level event extraction aims to extract structured event information
from unstructured text. However, a single document often contains limited event
information and the roles of different event arguments may be biased due to the
influence of the information source. This paper addresses the limitations of
tradit... | [
-0.5323576331,
-0.4084135592,
0.6388555169,
-0.1463189572,
-0.1977188438,
0.0507464372,
-0.2419230044,
-0.8383021355,
0.1651616544,
0.2425544709,
-0.5296382904,
-0.5770821571,
-0.8659616709,
0.0805325732,
-0.7472023964,
1.0858075619,
-0.0686886385,
-0.2724808455,
0.1190681905,
... |
http://arxiv.org/abs/2406.16020v2 | http://arxiv.org/pdf/2406.16020v2 | AudioBench: A Universal Benchmark for Audio Large Language Models | 2024-06-23T05:40:26 | 2024-06-25T12:27:06 | Bin Wang, Xunlong Zou, Geyu Lin, Shuo Sun, Zhuohan Liu, Wenyu Zhang, Zhengyuan Liu, AiTi Aw, Nancy F. Chen | cs.SD | cs.SD, cs.CL, eess.AS | We introduce AudioBench, a new benchmark designed to evaluate audio large
language models (AudioLLMs). AudioBench encompasses 8 distinct tasks and 26
carefully selected or newly curated datasets, focusing on speech understanding,
voice interpretation, and audio scene understanding. Despite the rapid
advancement of larg... | [
-0.5415734053,
-0.5792372823,
0.487306565,
0.4643853903,
0.147419408,
0.0102675231,
-0.471875608,
-0.5812316537,
-0.2241065651,
0.5219696164,
-0.5841248631,
-0.7783136368,
-0.3690973222,
-0.1245895177,
-0.7213152647,
0.5235517621,
0.2582852542,
0.3058237135,
-0.2830778062,
-0.4... |
http://arxiv.org/abs/2406.16013v1 | http://arxiv.org/pdf/2406.16013v1 | Database-Augmented Query Representation for Information Retrieval | 2024-06-23T05:02:21 | 2024-06-23T05:02:21 | Soyeong Jeong, Jinheon Baek, Sukmin Cho, Sung Ju Hwang, Jong C. Park | cs.CL | cs.CL, cs.AI, cs.IR | Information retrieval models that aim to search for the documents relevant to
the given query have shown many successes, which have been applied to diverse
tasks. However, the query provided by the user is oftentimes very short, which
challenges the retrievers to correctly fetch relevant documents. To tackle
this, exis... | [
-0.6833798885,
-0.7290825844,
0.3864578307,
-0.0847724974,
-0.3464384675,
0.4176827073,
0.2175178528,
-0.4666668177,
0.730191052,
0.7739149332,
-0.5383892655,
-0.7552419305,
-0.3087195754,
0.1698712409,
-0.4787545204,
0.5730339289,
0.3618556559,
-0.1458025426,
-0.3062120378,
-0... |
http://arxiv.org/abs/2406.16008v2 | http://arxiv.org/pdf/2406.16008v2 | Found in the Middle: Calibrating Positional Attention Bias Improves Long Context Utilization | 2024-06-23T04:35:42 | 2024-07-03T17:40:00 | Cheng-Yu Hsieh, Yung-Sung Chuang, Chun-Liang Li, Zifeng Wang, Long T. Le, Abhishek Kumar, James Glass, Alexander Ratner, Chen-Yu Lee, Ranjay Krishna, Tomas Pfister | cs.CL | cs.CL, cs.AI, cs.LG | Large language models (LLMs), even when specifically trained to process long
input contexts, struggle to capture relevant information located in the middle
of their input. This phenomenon has been known as the lost-in-the-middle
problem. In this work, we make three contributions. First, we set out to
understand the fac... | [
-0.500287354,
-0.5043211579,
1.1189604998,
0.2468405217,
-0.0992672667,
-0.0472211763,
-0.2857718468,
-0.0977293029,
0.0650228411,
0.4926542342,
-0.4348107874,
-0.1533521712,
-0.7516511679,
-0.1252026111,
-0.2689134777,
0.8141160607,
0.0216910876,
0.021530807,
-0.3387294114,
-0... |
http://arxiv.org/abs/2406.16007v1 | http://arxiv.org/pdf/2406.16007v1 | Distributed Rule Vectors is A Key Mechanism in Large Language Models' In-Context Learning | 2024-06-23T04:29:13 | 2024-06-23T04:29:13 | Bowen Zheng, Ming Ma, Zhongqiao Lin, Tianming Yang | cs.CL | cs.CL | Large Language Models (LLMs) have demonstrated remarkable abilities, one of
the most important being In-Context Learning (ICL). With ICL, LLMs can derive
the underlying rule from a few demonstrations and provide answers that comply
with the rule. Previous work hypothesized that the network creates a "task
vector" in sp... | [
-0.3836219013,
-0.8171823621,
0.4166343808,
0.30229792,
-0.1658527702,
-0.2003551871,
-0.1436518431,
-0.0810886696,
-0.0914464071,
0.3599917591,
-0.5851927996,
-0.6195719242,
-0.6425650716,
-0.0182603672,
-0.6249343157,
0.9665818214,
-0.0355212875,
-0.0622317158,
-0.5865786076,
... |
http://arxiv.org/abs/2406.15996v1 | http://arxiv.org/pdf/2406.15996v1 | Memorizing Documents with Guidance in Large Language Models | 2024-06-23T03:12:03 | 2024-06-23T03:12:03 | Bumjin Park, Jaesik Choi | cs.CL | cs.CL, cs.AI | Training data plays a pivotal role in AI models. Large language models (LLMs)
are trained with massive amounts of documents, and their parameters hold
document-related contents. Recently, several studies identified
content-specific locations in LLMs by examining the parameters. Instead of the
post hoc interpretation, w... | [
-0.0632212013,
-0.5427652001,
0.7109447122,
-0.2226949036,
-0.1496277601,
-0.3709692657,
-0.2729827464,
-0.0123620396,
-0.0155627271,
0.577883184,
-0.4906868041,
-0.6894284487,
-0.5012668967,
-0.469794631,
-0.5771811604,
0.9355623126,
-0.246592775,
0.2860937417,
-0.4325397909,
... |
http://arxiv.org/abs/2406.15992v1 | http://arxiv.org/pdf/2406.15992v1 | Can LLM Graph Reasoning Generalize beyond Pattern Memorization? | 2024-06-23T02:59:15 | 2024-06-23T02:59:15 | Yizhuo Zhang, Heng Wang, Shangbin Feng, Zhaoxuan Tan, Xiaochuang Han, Tianxing He, Yulia Tsvetkov | cs.CL | cs.CL, I.2.7 | Large language models (LLMs) demonstrate great potential for problems with
implicit graphical structures, while recent works seek to enhance the graph
reasoning capabilities of LLMs through specialized instruction tuning. The
resulting 'graph LLMs' are evaluated with in-distribution settings only, thus
it remains under... | [
-0.2331258059,
-1.0146043301,
0.8051480651,
-0.0514035746,
-0.2951342165,
0.4272241592,
-0.2788819671,
-0.4173237681,
-0.0041860049,
0.602159977,
-0.5593592525,
-0.9516939521,
-0.574439466,
0.2147143185,
-0.3089041412,
0.7017788291,
-0.1490241289,
-0.0802433491,
-0.2606519759,
... |
http://arxiv.org/abs/2406.15990v1 | http://arxiv.org/pdf/2406.15990v1 | Enhancing Cross-Document Event Coreference Resolution by Discourse Structure and Semantic Information | 2024-06-23T02:54:48 | 2024-06-23T02:54:48 | Qiang Gao, Bobo Li, Zixiang Meng, Yunlong Li, Jun Zhou, Fei Li, Chong Teng, Donghong Ji | cs.CL | cs.CL, cs.AI | Existing cross-document event coreference resolution models, which either
compute mention similarity directly or enhance mention representation by
extracting event arguments (such as location, time, agent, and patient),
lacking the ability to utilize document-level information. As a result, they
struggle to capture lon... | [
-0.2892920375,
-0.4668444395,
0.6237989068,
-0.1602597833,
-0.3367246687,
-0.0063639567,
-0.4310375452,
-0.7421108484,
0.3060311377,
0.1309903711,
-0.2635216713,
-0.6037163734,
-0.6695209146,
0.0709799603,
-0.1621157229,
0.9315009713,
-0.051981017,
-0.1129854247,
-0.0790804252,
... |
http://arxiv.org/abs/2406.15981v1 | http://arxiv.org/pdf/2406.15981v1 | Serial Position Effects of Large Language Models | 2024-06-23T02:02:52 | 2024-06-23T02:02:52 | Xiaobo Guo, Soroush Vosoughi | cs.CL | cs.CL | Large Language Models (LLMs) have shown remarkable capabilities in zero-shot
learning applications, generating responses to queries using only pre-training
information without the need for additional fine-tuning. This represents a
significant departure from traditional machine learning approaches. Previous
research has... | [
-0.2191980034,
-0.7736088634,
0.942448616,
0.0861018524,
-0.1170612201,
-0.2346713245,
-0.1503024399,
-0.3323559165,
0.235845238,
0.6430595517,
-0.3702107966,
-0.5990515947,
-0.8765252233,
0.0377935953,
-0.3748455346,
0.8361208439,
0.3182499409,
0.1290245652,
-0.1364234835,
-0.... |
http://arxiv.org/abs/2406.15968v1 | http://arxiv.org/pdf/2406.15968v1 | ReCaLL: Membership Inference via Relative Conditional Log-Likelihoods | 2024-06-23T00:23:13 | 2024-06-23T00:23:13 | Roy Xie, Junlin Wang, Ruomin Huang, Minxing Zhang, Rong Ge, Jian Pei, Neil Zhenqiang Gong, Bhuwan Dhingra | cs.CL | cs.CL, cs.LG | The rapid scaling of large language models (LLMs) has raised concerns about
the transparency and fair use of the pretraining data used for training them.
Detecting such content is challenging due to the scale of the data and limited
exposure of each instance during training. We propose ReCaLL (Relative
Conditional Log-... | [
-0.2554125488,
-0.4508451521,
0.2981896698,
-0.4062225521,
-0.1626706123,
-0.0059466711,
0.0289648436,
-0.338228941,
0.5519687533,
0.4506980777,
-0.7005785108,
-0.5970125794,
-0.5740349889,
0.1049259901,
-0.7101675868,
0.7369835973,
0.3499026895,
-0.1532107145,
-0.1345425695,
-... |
http://arxiv.org/abs/2406.15966v1 | http://arxiv.org/pdf/2406.15966v1 | Evaluating the Effectiveness of the Foundational Models for Q&A Classification in Mental Health care | 2024-06-23T00:11:07 | 2024-06-23T00:11:07 | Hassan Alhuzali, Ashwag Alasmari | cs.CL | cs.CL, cs.AI | Pre-trained Language Models (PLMs) have the potential to transform mental
health support by providing accessible and culturally sensitive resources.
However, despite this potential, their effectiveness in mental health care and
specifically for the Arabic language has not been extensively explored. To
bridge this gap, ... | [
-0.7916793823,
-0.8661454916,
0.6190162897,
-0.1115881503,
-0.2296570092,
-0.0058985525,
-0.1396811306,
-0.3654825389,
-0.2931358218,
0.3806808293,
-0.5098079443,
-0.6722160578,
-0.8203353286,
0.1000072435,
-0.3402912021,
1.1666973829,
-0.1054128259,
0.2015231997,
0.0973656252,
... |
http://arxiv.org/abs/2406.15963v1 | http://arxiv.org/pdf/2406.15963v1 | Effectiveness of ChatGPT in explaining complex medical reports to patients | 2024-06-23T00:04:07 | 2024-06-23T00:04:07 | Mengxuan Sun, Ehud Reiter, Anne E Kiltie, George Ramsay, Lisa Duncan, Peter Murchie, Rosalind Adam | cs.HC | cs.HC, cs.CL, q-bio.OT | Electronic health records contain detailed information about the medical
condition of patients, but they are difficult for patients to understand even
if they have access to them. We explore whether ChatGPT (GPT 4) can help
explain multidisciplinary team (MDT) reports to colorectal and prostate cancer
patients. These r... | [
-0.0691166073,
-0.6314560771,
0.9686900973,
0.203873381,
-0.3914955258,
0.3389233947,
-0.1747428328,
-0.6156713963,
0.100193508,
0.1955090463,
-0.5407260656,
-0.3743388057,
-1.0863435268,
0.1138382331,
-0.1050205082,
1.409375906,
-0.2002465874,
0.4864413142,
0.2962983847,
-0.04... |
http://arxiv.org/abs/2406.17806v1 | http://arxiv.org/pdf/2406.17806v1 | MOSSBench: Is Your Multimodal Language Model Oversensitive to Safe Queries? | 2024-06-22T23:26:07 | 2024-06-22T23:26:07 | Xirui Li, Hengguang Zhou, Ruochen Wang, Tianyi Zhou, Minhao Cheng, Cho-Jui Hsieh | cs.CL | cs.CL, cs.AI, cs.CR, cs.CV, cs.LG | Humans are prone to cognitive distortions -- biased thinking patterns that
lead to exaggerated responses to specific stimuli, albeit in very different
contexts. This paper demonstrates that advanced Multimodal Large Language
Models (MLLMs) exhibit similar tendencies. While these models are designed to
respond queries u... | [
-0.5613358021,
-0.9974043965,
0.4213125408,
0.383282721,
-0.0029046577,
-0.3464685082,
-0.0483002365,
-0.5881003737,
-0.3360388279,
0.2890229225,
-0.5296986103,
-0.0628047585,
-0.4542484581,
-0.1873324513,
-0.4611720741,
1.0164086819,
0.1904917359,
0.0616523102,
-0.104915455,
-... |
http://arxiv.org/abs/2406.17805v1 | http://arxiv.org/pdf/2406.17805v1 | Can LLMs Generate Visualizations with Dataless Prompts? | 2024-06-22T22:59:09 | 2024-06-22T22:59:09 | Darius Coelho, Harshit Barot, Naitik Rathod, Klaus Mueller | cs.CL | cs.CL, cs.AI, cs.HC | Recent advancements in large language models have revolutionized information
access, as these models harness data available on the web to address complex
queries, becoming the preferred information source for many users. In certain
cases, queries are about publicly available data, which can be effectively
answered with... | [
-0.1931452453,
-0.8342279196,
1.1309520006,
0.2534302771,
-0.0499524362,
-0.0051216632,
-0.0868523791,
-0.2341762036,
0.2507404983,
0.5383878946,
-0.7550932169,
-0.818877697,
-0.3080927432,
0.2095758915,
-0.5876575708,
0.9253786802,
0.019368533,
0.0519736595,
-0.4422336221,
0.0... |
http://arxiv.org/abs/2406.15951v1 | http://arxiv.org/pdf/2406.15951v1 | Modular Pluralism: Pluralistic Alignment via Multi-LLM Collaboration | 2024-06-22T22:07:40 | 2024-06-22T22:07:40 | Shangbin Feng, Taylor Sorensen, Yuhan Liu, Jillian Fisher, Chan Young Park, Yejin Choi, Yulia Tsvetkov | cs.CL | cs.CL | While existing alignment paradigms have been integral in developing large
language models (LLMs), LLMs often learn an averaged human preference and
struggle to model diverse preferences across cultures, demographics, and
communities. We propose Modular Pluralism, a modular framework based on
multi-LLM collaboration for... | [
-0.4821412265,
-0.4601102769,
0.6804878116,
0.291754216,
-0.0937786251,
0.4456843138,
-0.3366120756,
-0.1368458718,
0.6196886897,
0.4816889465,
-0.443990916,
-0.5143498778,
-0.4527347982,
0.0873390064,
0.0430041924,
0.7164137363,
0.1725619435,
0.2034196258,
-0.2811683714,
-0.45... |
http://arxiv.org/abs/2406.15948v1 | http://arxiv.org/pdf/2406.15948v1 | Teaching LLMs to Abstain across Languages via Multilingual Feedback | 2024-06-22T21:59:12 | 2024-06-22T21:59:12 | Shangbin Feng, Weijia Shi, Yike Wang, Wenxuan Ding, Orevaoghene Ahia, Shuyue Stella Li, Vidhisha Balachandran, Sunayana Sitaram, Yulia Tsvetkov | cs.CL | cs.CL | Multilingual LLMs often have knowledge disparities across languages, with
larger gaps in under-resourced languages. Teaching LLMs to abstain in the face
of knowledge gaps is thus a promising strategy to mitigate hallucinations in
multilingual settings. However, previous studies on LLM abstention primarily
focus on Engl... | [
-0.491481781,
-0.7058179379,
0.4672411084,
0.3100976944,
-0.1637444049,
0.4204206467,
-0.6129550338,
-0.5692029595,
0.2899580002,
0.4472465813,
-0.7250791788,
-0.545425415,
-0.3429742754,
0.2431838661,
0.1515728831,
0.920173645,
-0.274491787,
0.5046349168,
-0.3482649326,
-0.353... |
http://arxiv.org/abs/2406.15940v1 | http://arxiv.org/pdf/2406.15940v1 | Beyond Individual Facts: Investigating Categorical Knowledge Locality of Taxonomy and Meronomy Concepts in GPT Models | 2024-06-22T21:12:57 | 2024-06-22T21:12:57 | Christopher Burger, Yifan Hu, Thai Le | cs.LG | cs.LG, cs.AI, cs.CL | The location of knowledge within Generative Pre-trained Transformer
(GPT)-like models has seen extensive recent investigation. However, much of the
work is focused towards determining locations of individual facts, with the end
goal being the editing of facts that are outdated, erroneous, or otherwise
harmful, without ... | [
-0.6814009547,
-0.7878765464,
0.8271077275,
-0.3809049428,
-0.01282892,
-0.0848972425,
-0.0741067529,
-0.2230265141,
0.1343466491,
0.5120444894,
-0.41490224,
-0.5628696084,
-0.6302677393,
-0.1562119722,
-0.3226898313,
1.1740689278,
0.2252505124,
-0.3540510237,
-0.4446097314,
-0... |
http://arxiv.org/abs/2406.15938v1 | http://arxiv.org/pdf/2406.15938v1 | RuleR: Improving LLM Controllability by Rule-based Data Recycling | 2024-06-22T20:57:12 | 2024-06-22T20:57:12 | Ming Li, Han Chen, Chenguang Wang, Dang Nguyen, Dianqi Li, Tianyi Zhou | cs.CL | cs.CL, cs.AI, cs.LG | Large language models (LLMs) still lack delicate controllability over their
responses, which is critical to enhancing their performance and the user
experience. However, curating supervised fine-tuning (SFT) datasets to improve
LLM controllability usually relies on human experts or proprietary LLMs, which
requires addi... | [
-0.3522142172,
-0.6453573704,
0.2056959718,
0.0244754683,
-0.2160722911,
0.0059309169,
-0.4391383529,
-0.3077543676,
0.2065882683,
0.8912428021,
-0.5427469611,
-0.1913756579,
-0.540096283,
-0.0527313948,
-0.6142857075,
1.0248149633,
-0.2496137619,
-0.0827618837,
-0.3340599239,
... |
http://arxiv.org/abs/2406.15927v1 | http://arxiv.org/pdf/2406.15927v1 | Semantic Entropy Probes: Robust and Cheap Hallucination Detection in LLMs | 2024-06-22T19:46:06 | 2024-06-22T19:46:06 | Jannik Kossen, Jiatong Han, Muhammed Razzak, Lisa Schut, Shreshth Malik, Yarin Gal | cs.CL | cs.CL, cs.AI, cs.LG | We propose semantic entropy probes (SEPs), a cheap and reliable method for
uncertainty quantification in Large Language Models (LLMs). Hallucinations,
which are plausible-sounding but factually incorrect and arbitrary model
generations, present a major challenge to the practical adoption of LLMs.
Recent work by Farquha... | [
-0.3206827939,
-0.6123837829,
0.9415509105,
-0.0125396997,
-0.2826485932,
-0.3256797791,
-0.0652740747,
-0.6220893264,
0.394467175,
0.5527259707,
-0.7296911478,
-0.6249370575,
-0.6993635297,
0.1711416245,
-0.6539319158,
0.3260032833,
0.2126577795,
0.1654683799,
-0.1508009136,
0... |
http://arxiv.org/abs/2406.15891v1 | http://arxiv.org/pdf/2406.15891v1 | The Unlikely Duel: Evaluating Creative Writing in LLMs through a Unique Scenario | 2024-06-22T17:01:59 | 2024-06-22T17:01:59 | Carlos Gómez-Rodríguez, Paul Williams | cs.CL | cs.CL, 68T50, I.2.7 | This is a summary of the paper "A Confederacy of Models: a Comprehensive
Evaluation of LLMs on Creative Writing", which was published in Findings of
EMNLP 2023. We evaluate a range of recent state-of-the-art, instruction-tuned
large language models (LLMs) on an English creative writing task, and compare
them to human w... | [
-0.1699736267,
-0.7224704623,
0.8593123555,
-0.0647914037,
-0.0995758399,
0.5205333233,
-0.2467128932,
-0.3641323745,
0.4332093298,
0.488622874,
-0.4034524858,
-0.563226819,
-0.3759259284,
0.2147446573,
-0.205686152,
0.7403388619,
-0.2605981231,
-0.0131225735,
-0.3882026672,
-0... |
http://arxiv.org/abs/2406.15890v1 | http://arxiv.org/pdf/2406.15890v1 | Language Alignment via Nash-learning and Adaptive feedback | 2024-06-22T16:55:21 | 2024-06-22T16:55:21 | Ari Azarafrooz, Farshid Faal | cs.LG | cs.LG, cs.AI, cs.CL, cs.GT | Recent research has shown the potential of Nash Learning via Human Feedback
for large language model alignment by incorporating the notion of a preference
model in a minimax game setup. We take this idea further by casting the
alignment as a mirror descent algorithm against the adaptive feedback of an
improved opponent... | [
-0.1588951796,
-0.5157970786,
0.4405724406,
-0.1193271428,
0.1678714603,
0.3564930856,
-0.2757049501,
-0.1841170937,
0.4541266561,
0.4890941978,
-0.7628587484,
-0.2806289196,
-0.4511931837,
-0.0088442154,
-0.153406918,
1.2530961037,
-0.0633137748,
0.4090586603,
-0.1382217705,
-... |
http://arxiv.org/abs/2406.15888v1 | http://arxiv.org/pdf/2406.15888v1 | Real-time Speech Summarization for Medical Conversations | 2024-06-22T16:37:51 | 2024-06-22T16:37:51 | Khai Le-Duc, Khai-Nguyen Nguyen, Long Vo-Dang, Truong-Son Hy | cs.CL | cs.CL, cs.AI, cs.LG, cs.SD, eess.AS | In doctor-patient conversations, identifying medically relevant information
is crucial, posing the need for conversation summarization. In this work, we
propose the first deployable real-time speech summarization system for
real-world applications in industry, which generates a local summary after
every N speech uttera... | [
-0.1352346539,
-0.7819503546,
0.6209844351,
0.2855574191,
-0.4415707886,
0.3531039655,
0.0595385619,
-0.8945576549,
0.871132493,
0.5720295906,
-0.5457257628,
-0.6437726617,
-0.4228625298,
-0.1418187767,
-0.3876358569,
0.9108194709,
0.3141945899,
0.0614071712,
-0.1997433007,
-0.... |
http://arxiv.org/abs/2406.15883v1 | http://arxiv.org/pdf/2406.15883v1 | SimSMoE: Solving Representational Collapse via Similarity Measure | 2024-06-22T16:10:45 | 2024-06-22T16:10:45 | Giang Do, Hung Le, Truyen Tran | cs.CL | cs.CL, cs.AI | Sparse mixture of experts (SMoE) have emerged as an effective approach for
scaling large language models while keeping a constant computational cost.
Regardless of several notable successes of SMoE, effective training such
architecture remains elusive due to the representation collapse problem, which
in turn harms mode... | [
-0.7332097292,
-0.0581735186,
0.463323772,
0.0315612517,
-0.141240418,
0.0542389229,
-0.4829545319,
-0.0298555028,
0.6975420713,
0.4930343926,
-0.7037007213,
-0.1548531801,
-0.7469267249,
-0.0998786092,
-0.1226767898,
0.4446791708,
0.1422899514,
0.0881935656,
-0.4186028242,
-0.... |
http://arxiv.org/abs/2406.15877v2 | http://arxiv.org/pdf/2406.15877v2 | BigCodeBench: Benchmarking Code Generation with Diverse Function Calls and Complex Instructions | 2024-06-22T15:52:04 | 2024-06-26T17:05:14 | Terry Yue Zhuo, Minh Chien Vu, Jenny Chim, Han Hu, Wenhao Yu, Ratnadira Widyasari, Imam Nur Bani Yusuf, Haolan Zhan, Junda He, Indraneil Paul, Simon Brunner, Chen Gong, Thong Hoang, Armel Randy Zebaze, Xiaoheng Hong, Wen-Ding Li, Jean Kaddour, Ming Xu, Zhihan Zhang, Prateek Yadav, Naman Jain, Alex Gu, Zhoujun Cheng, Ji... | cs.SE | cs.SE, cs.AI, cs.CL | Automated software engineering has been greatly empowered by the recent
advances in Large Language Models (LLMs) for programming. While current
benchmarks have shown that LLMs can perform various software engineering tasks
like human developers, the majority of their evaluations are limited to short
and self-contained ... | [
-0.550237,
-0.7174447179,
0.5613750219,
0.4953155518,
-0.0276544373,
0.2422189116,
-0.4300347865,
-0.5582565665,
-0.4432605207,
0.4643746316,
-0.5912246108,
-0.7533511519,
-0.4557779133,
0.2195300311,
-0.4612027705,
0.584028542,
-0.2984493077,
0.2333123982,
-0.2835907042,
-0.07... |
http://arxiv.org/abs/2406.15871v1 | http://arxiv.org/pdf/2406.15871v1 | Uncovering Hidden Intentions: Exploring Prompt Recovery for Deeper Insights into Generated Texts | 2024-06-22T15:16:11 | 2024-06-22T15:16:11 | Louis Give, Timo Zaoral, Maria Antonietta Bruno | cs.CL | cs.CL, cs.AI | Today, the detection of AI-generated content is receiving more and more
attention. Our idea is to go beyond detection and try to recover the prompt
used to generate a text. This paper, to the best of our knowledge, introduces
the first investigation in this particular domain without a closed set of
tasks. Our goal is t... | [
-0.1690395623,
-0.9834052324,
1.045638442,
-0.3192780614,
-0.2171919048,
-0.1890305132,
0.0596035384,
-0.4453896284,
0.3124338388,
0.7235534191,
-1.0947397947,
-0.4710459709,
-0.5124320984,
0.3597466946,
-0.5249248147,
1.1990655661,
0.1194733977,
-0.3739706278,
0.1859921515,
0.... |
http://arxiv.org/abs/2406.15869v1 | http://arxiv.org/pdf/2406.15869v1 | A multitask learning framework for leveraging subjectivity of annotators to identify misogyny | 2024-06-22T15:06:08 | 2024-06-22T15:06:08 | Jason Angel, Segun Taofeek Aroyehun, Grigori Sidorov, Alexander Gelbukh | cs.CL | cs.CL | Identifying misogyny using artificial intelligence is a form of combating
online toxicity against women. However, the subjective nature of interpreting
misogyny poses a significant challenge to model the phenomenon. In this paper,
we propose a multitask learning approach that leverages the subjectivity of
this task to ... | [
0.0716742724,
-0.7023332119,
0.4891842008,
0.3807096183,
-0.2826411724,
0.2546226382,
-0.1364294142,
-0.382950753,
0.0179760754,
0.3744489849,
-0.7303192019,
-0.3752419949,
-0.8284175396,
0.3664325476,
-0.2976127863,
1.143025279,
0.0717173219,
0.1268123984,
-0.2260311693,
-0.04... |
http://arxiv.org/abs/2406.17803v1 | http://arxiv.org/pdf/2406.17803v1 | Understanding the Role of User Profile in the Personalization of Large Language Models | 2024-06-22T14:32:35 | 2024-06-22T14:32:35 | Bin Wu, Zhengyan Shi, Hossein A. Rahmani, Varsha Ramineni, Emine Yilmaz | cs.CL | cs.CL, cs.AI, cs.IR | Utilizing user profiles to personalize Large Language Models (LLMs) has been
shown to enhance the performance on a wide range of tasks. However, the precise
role of user profiles and their effect mechanism on LLMs remains unclear. This
study first confirms that the effectiveness of user profiles is primarily due
to per... | [
-0.5527320504,
-0.6239212155,
0.8609510064,
0.2002618462,
0.2852143645,
0.2502835393,
-0.2825514078,
-0.5119259357,
0.2619021237,
0.8796797395,
-0.3731843531,
-0.5067815781,
-0.3945670128,
0.0673453286,
-0.3233136535,
0.8986824155,
0.2294735312,
-0.1276686192,
-0.3214868605,
0.... |
http://arxiv.org/abs/2406.15862v1 | http://arxiv.org/pdf/2406.15862v1 | Speech Analysis of Language Varieties in Italy | 2024-06-22T14:19:51 | 2024-06-22T14:19:51 | Moreno La Quatra, Alkis Koudounas, Elena Baralis, Sabato Marco Siniscalchi | cs.CL | cs.CL | Italy exhibits rich linguistic diversity across its territory due to the
distinct regional languages spoken in different areas. Recent advances in
self-supervised learning provide new opportunities to analyze Italy's
linguistic varieties using speech data alone. This includes the potential to
leverage representations l... | [
-0.6487295628,
-0.4933970869,
0.3011056781,
0.3050769269,
-0.0219931137,
-0.0712535754,
-0.4340760708,
-0.4353632629,
0.4197202325,
0.506501317,
-0.4491741657,
-0.6046839356,
-0.4852529466,
0.191917032,
-0.2786597311,
0.9014853835,
0.3243057132,
0.4400295019,
0.3786016405,
-0.3... |
http://arxiv.org/abs/2406.15846v1 | http://arxiv.org/pdf/2406.15846v1 | Revisiting Interpolation Augmentation for Speech-to-Text Generation | 2024-06-22T13:24:58 | 2024-06-22T13:24:58 | Chen Xu, Jie Wang, Xiaoqian Liu, Qianqian Dong, Chunliang Zhang, Tong Xiao, Jingbo Zhu, Dapeng Man, Wu Yang | cs.CL | cs.CL, eess.AS | Speech-to-text (S2T) generation systems frequently face challenges in
low-resource scenarios, primarily due to the lack of extensive labeled
datasets. One emerging solution is constructing virtual training samples by
interpolating inputs and labels, which has notably enhanced system
generalization in other domains. Des... | [
-0.560454905,
-0.4160225391,
0.3560601473,
0.6121877432,
-0.3453953862,
0.4142654836,
-0.2666855752,
-0.6062627435,
0.0102101211,
0.2425189316,
-0.743858397,
0.1875903159,
-0.7799091935,
0.1769921482,
-0.5464571118,
1.0107996464,
0.2641564012,
-0.1146844104,
-0.5863339305,
0.11... |
http://arxiv.org/abs/2407.02514v1 | http://arxiv.org/pdf/2407.02514v1 | LOGIC-LM++: Multi-Step Refinement for Symbolic Formulations | 2024-06-22T12:50:41 | 2024-06-22T12:50:41 | Shashank Kirtania, Priyanshu Gupta, Arjun Radhakirshna | cs.LO | cs.LO, cs.AI, cs.CL | In this paper we examine the limitations of Large Language Models (LLMs) for
complex reasoning tasks. While current approaches leverage formal languages as
intermediate representation of reasoning problems, they struggle with
generating intermediate formal specifications and refining these
representations. To address t... | [
-0.3380715251,
-0.8914966583,
0.9248824716,
0.1610822082,
0.0800608322,
0.400307864,
-0.3083338439,
-0.8447507024,
-0.0854743049,
0.5941547751,
-0.7618514299,
-0.3490970135,
-0.4973199666,
0.43890661,
-0.1910125166,
0.459926784,
-0.419465363,
0.4438186884,
-0.1447135508,
-0.204... |
http://arxiv.org/abs/2406.15823v1 | http://arxiv.org/pdf/2406.15823v1 | CaT-BENCH: Benchmarking Language Model Understanding of Causal and Temporal Dependencies in Plans | 2024-06-22T11:46:04 | 2024-06-22T11:46:04 | Yash Kumar Lal, Vanya Cohen, Nathanael Chambers, Niranjan Balasubramanian, Raymond Mooney | cs.CL | cs.CL | Understanding the abilities of LLMs to reason about natural language plans,
such as instructional text and recipes, is critical to reliably using them in
decision-making systems. A fundamental aspect of plans is the temporal order in
which their steps needs to be executed, which reflects the underlying causal
dependenc... | [
-0.2846386433,
-0.7180078626,
0.9381989241,
0.6837953925,
0.0952912793,
-0.0085662361,
0.1805554181,
-0.4081099927,
0.0381597653,
0.6242551208,
-0.7567178011,
-0.4123195112,
-0.4776783586,
0.0600285381,
-0.2223954201,
0.3774922788,
-0.3406162858,
0.5005378723,
-0.3522606194,
-0... |
http://arxiv.org/abs/2406.17801v1 | http://arxiv.org/pdf/2406.17801v1 | A multi-speaker multi-lingual voice cloning system based on vits2 for limmits 2024 challenge | 2024-06-22T10:49:36 | 2024-06-22T10:49:36 | Xiaopeng Wang, Yi Lu, Xin Qi, Zhiyong Wang, Yuankun Xie, Shuchen Shi, Ruibo Fu | cs.SD | cs.SD, cs.CL, eess.AS | This paper presents the development of a speech synthesis system for the
LIMMITS'24 Challenge, focusing primarily on Track 2. The objective of the
challenge is to establish a multi-speaker, multi-lingual Indic Text-to-Speech
system with voice cloning capabilities, covering seven Indian languages with
both male and fema... | [
-0.3256367743,
-0.3345639408,
0.0836352184,
0.6243662238,
-0.377517581,
0.3855082095,
-0.1000250727,
-0.5074138045,
0.3000296652,
0.4690658748,
-0.8883516788,
0.0628333986,
0.0448546372,
0.2748691738,
-0.3859883547,
0.9691982269,
0.3846349418,
0.1316804886,
0.0725153834,
-0.027... |
http://arxiv.org/abs/2406.15812v1 | http://arxiv.org/pdf/2406.15812v1 | Intrinsic Dimension Correlation: uncovering nonlinear connections in multimodal representations | 2024-06-22T10:36:04 | 2024-06-22T10:36:04 | Lorenzo Basile, Santiago Acevedo, Luca Bortolussi, Fabio Anselmi, Alex Rodriguez | cs.LG | cs.LG, cs.CL, cs.CV | To gain insight into the mechanisms behind machine learning methods, it is
crucial to establish connections among the features describing data points.
However, these correlations often exhibit a high-dimensional and strongly
nonlinear nature, which makes them challenging to detect using standard
methods. This paper exp... | [
-0.4129229784,
-1.0010544062,
0.4364567399,
0.2095254511,
0.0366215967,
0.1266065389,
0.1257336736,
-0.4107777178,
0.6368381977,
0.1223630682,
-0.2074963301,
-0.1344068497,
-0.5229626894,
-0.1613684595,
-0.3755988777,
0.6790593863,
-0.1522053331,
-0.0236482527,
-0.8528510928,
-... |
http://arxiv.org/abs/2406.15809v1 | http://arxiv.org/pdf/2406.15809v1 | LaMSUM: A Novel Framework for Extractive Summarization of User Generated Content using LLMs | 2024-06-22T10:25:55 | 2024-06-22T10:25:55 | Garima Chhikara, Anurag Sharma, V. Gurucharan, Kripabandhu Ghosh, Abhijnan Chakraborty | cs.CL | cs.CL, cs.LG | Large Language Models (LLMs) have demonstrated impressive performance across
a wide range of NLP tasks, including summarization. Inherently LLMs produce
abstractive summaries, and the task of achieving extractive summaries through
LLMs still remains largely unexplored. To bridge this gap, in this work, we
propose a nov... | [
-0.3225021362,
-0.4183451235,
0.5240783691,
0.3246801794,
-0.6053841114,
0.0041472786,
-0.2814136147,
-0.200212881,
0.6783087254,
0.9425879717,
-0.4279386401,
-0.7687131763,
-0.6892357469,
0.1425920129,
-0.7366667986,
1.0140357018,
0.0382608697,
-0.3018947542,
-0.2180910259,
0.... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.