Chunk-Level Guided Generation uses off-the-shelf large LLMs to score fixed-length chunks from small models via likelihoods, matching trained PRM performance on math benchmarks without reward-model training.
arXiv preprint arXiv:2405.19262 , year=
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
representative citing papers
PRIME enables online process reward model updates in LLM RL using implicit rewards from rollouts and outcome labels, yielding 15.1% average gains on reasoning benchmarks and surpassing a stronger instruct model with 10% of the data.
Trust functions filter unreliable weak labels to enable near-lossless weak-to-strong generalization and iterative chaining.
citing papers explorer
-
Off-the-Shelf LLMs as Process Scorers: Training-Free Alternative to PRMs for Mathematical Reasoning
Chunk-Level Guided Generation uses off-the-shelf large LLMs to score fixed-length chunks from small models via likelihoods, matching trained PRM performance on math benchmarks without reward-model training.
-
Process Reinforcement through Implicit Rewards
PRIME enables online process reward model updates in LLM RL using implicit rewards from rollouts and outcome labels, yielding 15.1% average gains on reasoning benchmarks and surpassing a stronger instruct model with 10% of the data.
-
Trust Functions: Near-Lossless Weak-to-Strong Generalization by Learning When to Trust the Weak Teacher
Trust functions filter unreliable weak labels to enable near-lossless weak-to-strong generalization and iterative chaining.