A*-decoding applies A* search with a PRM heuristic to LLM decoding, improving accuracy per generated token on MATH500 and AIME 2024 compared to best-of-N and particle filtering.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.AI 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
A*-Decoding: Token-Efficient Inference Scaling
A*-decoding applies A* search with a PRM heuristic to LLM decoding, improving accuracy per generated token on MATH500 and AIME 2024 compared to best-of-N and particle filtering.