Dynamically adjusting beta via LLM-as-judge downweights biased comparisons to learn more rational reward models from flawed human preferences.
Proceedings of the 2026 ACM Conference on Fairness, Accountability, and Transparency (FAccT '26) , year =
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.AI 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Mitigating Cognitive Bias in RLHF by Altering Rationality
Dynamically adjusting beta via LLM-as-judge downweights biased comparisons to learn more rational reward models from flawed human preferences.