Papers2

All Beginner Intermediate Advanced

All Sources arXiv

#MATH500

Beyond Hard Masks: Progressive Token Evolution for Diffusion Language Models

Intermediate

Linhao Zhong, Linyu Wu et al.Jan 12arXiv

Diffusion Language Models (DLMs) write by polishing whole sentences in several passes instead of one token at a time.

#Diffusion Language Models#Masked Diffusion#Soft Token Distributions

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