bias-measurement

Tag

Cards List
#bias-measurement

Every Token Counts: Exact Likert-Scale Distributions for Measuring LLM Attitudes and Biases

arXiv cs.CL · 2026-08-12 Cached

Introduces an analytically exact framework for controlled behavioral evaluation of LLMs, using fully crossed factorial experiments and exact token-level probability mass functions to isolate causal biases that aggregate benchmarks obscure.

0 favorites 0 likes
#bias-measurement

Calibrating WEAT Against Anisotropy: ZCA Whitening as a Geometric Pre-Processing Step for Embedding Association Tests

arXiv cs.CL · 2026-08-10 Cached

This paper proposes ZCA whitening as a geometric pre-processing step for WEAT to address embedding anisotropy, showing that calibration changes significance status for over 30% of results and that uncalibrated bias measurements may be unreliable.

0 favorites 0 likes
← Back to home

Submit Feedback