A single process-level reward model, trained on weakly labeled biased versus unbiased reasoning, transfers across tasks and models to reduce equalized odds gaps in LLM decision-making.
On the dangers of stochastic parrots: Can language models be too big? In Proceedings of the 2021 ACM conference on fairness, accountability, and transparency, 2021
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Guiding LLM Decision-Making with Fairness Reward Models
A single process-level reward model, trained on weakly labeled biased versus unbiased reasoning, transfers across tasks and models to reduce equalized odds gaps in LLM decision-making.