Entity · technique

Faithfulness Evaluation

techniqueactivefaithfulness-evaluation-d4d914c2·1 events·first seen Jun 1, 2026

Aliases: Faithfulness Evaluation

Co-occurring entities

Token-level Rationales Plausibility Evaluation hate speech detection Soft Label Supervision

More like this (12)

Relaxing Faithfulness with Intervention-Only Causal Discovery Faithfulness to Refusal: A Causal Audit of Neuron Selectors intervention-immediacy faithfulness Quantifying Faithful Confidence Expression in Large Reasoning Models Plausibility Evaluation Faith-Shap EDIT: Evidence-Diagnosed Intervention Training for Rule-Faithful LLM Grading BeliefTrack Fidelity Management & Research Company SycophancyEval Model Forensics: Investigating Whether Concerning Behavior Reflects Misalignment L-Eval

Recent events (1)

5arXiv · cs.CL·Jun 1, 2026·source ↗

Disagreeing Rationales: Rethinking Classification and Explainability Evaluation in Hate Speech Detection

This paper investigates human disagreement in token-level rationale annotations for hate speech detection, a dimension less studied than label disagreement. The authors unify diverse models, training strategies, loss functions, and evaluation metrics under a single protocol, systematically comparing hard and soft label/rationale representation spaces. Results show that both hard and soft metrics favor softer representations, suggesting that soft supervision better captures human reasoning variation in subjective NLP tasks. The work calls for rethinking evaluation frameworks for classification and explainability in subjective NLP.

Evaluation and Benchmarking Alignment and RLHF Token-level Rationales Faithfulness Evaluation Plausibility Evaluation +2 more