ARCA: Adapter-Residual Credit Assignment When Token Signals Degenerate 事件

ARCA: Adapter-Residual Credit Assignment When Token Signals Degenerate arXiv:2606.00257v1 Announce Type: cross Abstract: Token-level credit assignment for language-model reinforcement learning is usually formulated as if the policy were fully trainable, while practical LLM-RL pipelines often rely on parameter-efficient fine-tuning, especially LoRA. We argue that this separation hides a structural failure mode. Under LoRA, the policy is restricted to a low-rank neighborhood of the reference mode