Papers2

#DeepScaleR

Likelihood-Based Reward Designs for General LLM Reasoning

Ariel Kwiatkowski, Natasha Butt et al.Feb 3arXiv

Binary right/wrong rewards for training reasoning in large language models are hard to design and often too sparse to learn from.

#log-likelihood reward#chain-of-thought (CoT)#reinforcement learning for LLMs

Exploration vs Exploitation: Rethinking RLVR through Clipping, Entropy, and Spurious Reward

Intermediate

Peter Chen, Xiaopeng Li et al.Dec 18arXiv

The paper studies why two opposite-sounding tricks in RL for reasoning—adding random (spurious) rewards and reducing randomness (entropy)—can both seem to help large language models think better.

#RLVR#Group Relative Policy Optimization#ratio clipping