MétaCan
Menu
Back to cohort

Hard Negatives, Hard Lessons: Revisiting Training Data Quality for Robust Information Retrieval with LLMs

2025· article· W4416034664 on OpenAlexfundno aff
Nandan Thakur, Crystina Zhang, Xueguang Ma, Jimmy Lin

Bibliographic record

Venuenot available
Typearticle
Language
FieldDecision Sciences
TopicData Quality and Management
Canadian institutionsnot available
FundersNatural Sciences and Engineering Research Council of Canada
KeywordsTraining (meteorology)Quality (philosophy)Data qualityData retrievalTraining setInformation system

Abstract

fetched live from OpenAlex

Training robust retrieval and reranker models typically relies on large-scale retrieval datasets; for example, the BGE collection contains 1.6 million query-passage pairs sourced from various data sources.However, we find that certain datasets can negatively impact model effectiveness -pruning 8 out of 15 datasets from the BGE collection, reduces the training set size by 2.35×, surprisingly increases nDCG@10 on BEIR by 1.0 point.This motivates a deeper examination of training data quality, with a particular focus on "false negatives", where relevant passages are incorrectly labeled as irrelevant.We utilize LLMs as a simple, cost-effective approach to identify and relabel false negatives in training datasets.Experimental results show that relabeling false negatives as true positives improves both E5 (base) and Qwen2.5-7Bretrieval models by 0.7-1.4 points on BEIR and by 1.7-1.8 points at nDCG@10 on zero-shot AIR-BENCH evaluation.Similar gains are observed for rerankers fine-tuned on the relabeled data, such as Qwen2.5-3B on BEIR.The reliability of LLMs to identify false negatives is supported by human annotation results.Our training dataset and code are publicly available.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.

How this classification was reachedexpand

Full frame machine prediction

Teacher imitation

Not calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.

metaresearch head score (Codex)0.069
metaresearch head score (Gemma)0.375
Version: metacan-v3-hybrid-931329e0061cValidation status: machine_predicted_unvalidated
Candidate categoriesnone
Consensus categoriesnone
DomainCandidate signal: none · Consensus signal: none
Study designCandidate signal: Not applicable · Consensus signal: none
GenreCandidate signal: Empirical · Consensus signal: none
Teacher disagreement score0.069
Threshold uncertainty score0.363

Distilled classifier scores by category (both heads)

CategoryCodexGemma
Metaresearch0.0690.375
Meta-epidemiology (narrow)0.0010.001
Meta-epidemiology (broad)0.0030.001
Bibliometrics0.0040.003
Science and technology studies0.0020.005
Scholarly communication0.0080.015
Open science0.0050.007
Research integrity0.0030.006
Insufficient payload (model declined to judge)0.0040.001

Machine scores (provisional)

The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.

Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.

Opus teacher head0.571
GPT teacher head0.486
Teacher spread0.085 · how far apart the two teachers sit on this one work
Validation statusscore_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it

Classification

machine, unvalidated

Machine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.

The models applied no category: nothing in the taxonomy fit this work.
Study designNot applicable
Domainnot available
GenreEmpirical

How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".

Quick stats

Citations1
Published2025
Admission routes1
Has abstractno

Explore more

Same topicData Quality and ManagementFrench-language works237,207