Adding a batch-wise sum-to-zero penalty to Bradley-Terry reward modeling makes reward models more robust to unseen prompts and responses, according to experiments across multiple model families and benchmarks.
Back to basics: Revisiting REINFORCE -style optimization for learning from human feedback in LLM s
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
On the Robustness of Reward Models for Language Model Alignment
Adding a batch-wise sum-to-zero penalty to Bradley-Terry reward modeling makes reward models more robust to unseen prompts and responses, according to experiments across multiple model families and benchmarks.