Datasets:
id stringlengths 2 4 | text stringlengths 2.21k 7.34k | n_tokens int32 462 1.29k | embedding listlengths 59.1k 165k |
|---|---|---|---|
p0 | "Hereditary Cancer Syndromes, including Hereditary Breast and Ovarian Cancer (HBOC) and Lynch Syndro(...TRUNCATED) | 853 | [-0.0301055908203125,0.032958984375,0.0938720703125,-0.005466461181640625,-0.0188446044921875,0.0839(...TRUNCATED) |
p1 | "Hybrid F1 mice were injected with urethane (300 mg/kg) at 8 days of age. A group was then put on a (...TRUNCATED) | 884 | [0.083251953125,0.11358642578125,-0.0924072265625,0.1317138671875,0.02447509765625,0.10968017578125,(...TRUNCATED) |
p2 | "Decompressive craniectomy is a valid neurosurgical strategy now a day as an alternative to control (...TRUNCATED) | 833 | [0.0205078125,0.123291015625,-0.1611328125,0.022186279296875,0.0274658203125,-0.0010995864868164062,(...TRUNCATED) |
p3 | "disease. Early diagnosis ensures early management of the breast cancer and decreases the mortality (...TRUNCATED) | 774 | [0.0247955322265625,0.129150390625,0.053741455078125,0.0665283203125,0.0850830078125,0.19140625,-0.0(...TRUNCATED) |
p4 | "The cellular prion protein (PrP C ) is expressed as a cell surface protein mainly in the central an(...TRUNCATED) | 757 | [0.056365966796875,0.053253173828125,0.09271240234375,-0.0645751953125,-0.0931396484375,-0.040130615(...TRUNCATED) |
p5 | "FMF is considered as an autosomal recessive hereditary disease, associated with a single gene named(...TRUNCATED) | 897 | [0.01654052734375,0.045135498046875,0.00439453125,-0.06951904296875,0.095947265625,0.07281494140625,(...TRUNCATED) |
p6 | "The percentage of hypochromic red cells also showed a significant negative correlation with the CHr(...TRUNCATED) | 902 | [-0.08990478515625,-0.029571533203125,-0.06536865234375,0.04644775390625,-0.0055084228515625,-0.1295(...TRUNCATED) |
p7 | "There is also a need for further evidence supporting the doseeresponse relationships for many of th(...TRUNCATED) | 858 | [-0.004970550537109375,0.06365966796875,-0.0810546875,0.052734375,-0.150146484375,0.0740966796875,-0(...TRUNCATED) |
p8 | "I t is unequivocal that peristomal complications are one of the most challenging aspects of living (...TRUNCATED) | 895 | [-0.13037109375,-0.0287017822265625,-0.09759521484375,-0.0716552734375,0.01514434814453125,-0.014381(...TRUNCATED) |
p9 | "Osteomyelitis is an inflammatory process accompanied by bone destruction and caused by an infecting(...TRUNCATED) | 860 | [0.08734130859375,-0.0035228729248046875,0.037689208984375,-0.00791168212890625,-0.08636474609375,-0(...TRUNCATED) |
End of preview. Expand in Data Studio
MIRIAD 200, encoded with mLateOn-medical
Multi-vector (ColBERT-style) embeddings for
tomaarsen/miriad-benchmark-200k,
produced with multi-vector-encoder/mLateOn-medical.
| passages | 200 |
| token vectors | 176,014 |
| mean vectors / passage | 880.07 |
| dim | 128 |
| stored dtype | float16 |
| embeddings size | 0.05 GB |
| raw text encoded | 1 MB |
The embeddings are 49x larger than the text they came from, which is why late-interaction retrieval needs quantization or pooling.
Usage
embedding is flattened; reshape it to recover one vector per token.
import numpy as np
from datasets import load_dataset
ds = load_dataset("KShivendu/miriad-mlateon-colbert-smoke", split="train")
row = ds[0]
vecs = np.asarray(row["embedding"], dtype=np.float16).reshape(row["n_tokens"], 128)
MaxSim, the late-interaction score:
score = (query_vecs @ doc_vecs.T).max(axis=1).sum()
Queries, qrels and the raw layout live alongside in queries.npy, queries_offs.npy
and qrels.json.
- Downloads last month
- 42