Pith. sign in

REVIEW 13 cited by

On Measuring Social Biases in Sentence Encoders

Not yet reviewed by Pith; the record is open.

Re-run · record.json Download PDF Read on arXiv ↗

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 1903.10561 v1 pith:RP3D6C44 submitted 2019-03-25 cs.CL cs.CY

Chandler May , Alex Wang , Shikha Bordia , Samuel R. Bowman , Rachel Rudinger This is my paper · ORCID
classification cs.CLcs.CY
keywords encoderssentencetestbiaseswordsocialassociationbias
verification ladder T0 review T1 audit T2 compute T3 formal

Signed reviews

No signed human review yet.

0 comments
read the original abstract

The Word Embedding Association Test shows that GloVe and word2vec word embeddings exhibit human-like implicit biases based on gender, race, and other social constructs (Caliskan et al., 2017). Meanwhile, research on learning reusable text representations has begun to explore sentence-level texts, with some sentence encoders seeing enthusiastic adoption. Accordingly, we extend the Word Embedding Association Test to measure bias in sentence encoders. We then test several sentence encoders, including state-of-the-art methods such as ELMo and BERT, for the social biases studied in prior work and two important biases that are difficult or impossible to test at the word level. We observe mixed results including suspicious patterns of sensitivity that suggest the test's assumptions may not hold in general. We conclude by proposing directions for future work on measuring bias in sentence encoders.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 13 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. On The Effectiveness-Fluency Trade-Off In LLM Conditioning: A Systematic Study

    cs.CL 2026-06 unverdicted novelty 6.0 of 10

    Systematic experiments reveal that activation steering trades fluency for concept control, is less effective on instruction-tuned models, and that prompting/SFT excel at injection but not removal, with textual metrics...

  2. Position is Power: System Prompts as a Mechanism of Bias in Large Language Models (LLMs)

    cs.CY 2025-05 reject novelty 6.0 of 10

    Placing demographic audience information in system prompts rather than user prompts shifts sentiment and ranking outputs across six commercial LLMs, but the design confounds position with instruction content.

  3. Evaluate Bias without Manual Test Sets: A Concept Representation Perspective for LLMs

    cs.CL 2025-05 conditional novelty 6.0 of 10

    BiasLens uses concept activation vectors and sparse autoencoders to estimate LLM bias from internal representations, reporting moderate to strong agreement with behavioral bias metrics in a small evaluation.

  4. Mitigating Gender Bias in Contextual Word Embeddings

    cs.CL 2024-11 reject novelty 6.0 of 10

    Regularized masked-language modeling and name-masking reduce gender bias in embeddings, but the contextual results rely heavily on evaluation metrics aligned with the training objective.

  5. Good Secretaries, Bad Truck Drivers? Occupational Gender Stereotypes in Sentiment Analysis

    cs.CL 2019-06 unverdicted novelty 6.0 of 10

    Authors release a new 800-sentence gender-balanced profession dataset and use it to test occupational gender stereotypes in three sentiment analysis models.

  6. DebiasRAG: A Tuning-Free Path to Fair Generation in Large Language Models through Retrieval-Augmented Generation

    cs.CL 2026-05 unverdicted novelty 5.0 of 10

    DebiasRAG uses a three-stage RAG process to generate and rerank query-specific debiasing contexts that act as fairness constraints for LLM outputs.

  7. What is in a name? Mitigating Name Bias in Text Embeddings via Anonymization

    cs.CL 2025-02 conditional novelty 5.0 of 10

    Text-embedding models are biased by names, and removing names before embedding improves semantic similarity judgments on two evaluation tasks.

  8. Implicit Priors Editing in Stable Diffusion via Targeted Token Adjustment

    cs.CV 2024-12 conditional novelty 5.0 of 10

    EMBEDIT edits a single word token embedding in Stable Diffusion to steer implicit visual priors (e.g., making 'bear' generate 'polar bear'), reporting better accuracy than cross-attention editing while using far fewer...

  9. Private, Verifiable, and Auditable AI Systems

    cs.CR 2025-08 conditional novelty 4.0 of 10

    A thesis demonstrating partial prototypes for zk-verifiable model evaluation and privacy-preserving retrieval, and arguing these pieces can compose into end-to-end auditable AI systems.

  10. Navigating the Risks: A Survey of Security, Privacy, and Ethics Threats in LLM-Based Agents

    cs.AI 2024-11 conditional novelty 4.0 of 10

    A survey proposing a source-and-impact taxonomy (input, model, combined; security, privacy, ethics) for threats to LLM-based agents, with feature analysis and four case studies.

  11. A Survey on Bias and Fairness in Machine Learning

    cs.LG 2019-08 conditional novelty 3.0 of 10

    This survey catalogs types of bias, fairness definitions, and mitigation strategies across ML domains.

  12. Bias in Large Language Models: Origin, Evaluation, and Mitigation

    cs.CL 2024-11 unverdicted novelty 2.0 of 10

    A literature review that categorizes bias in LLMs, surveys evaluation and mitigation techniques, and discusses ethical implications.

  13. Improving LLM Group Fairness on Tabular Data via In-Context Learning

    cs.LG 2024-12

Pith tools