Composition-RL improves LLM reasoning by composing multiple verifiable prompts into new training questions for RL.
Since AIME24 and AIME25 each contain 30 problems, we report pass@1 using 32 samples per problem (avg@32)
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Composition-RL: Compose Your Verifiable Prompts for Reinforcement Learning of Large Language Models
Composition-RL improves LLM reasoning by composing multiple verifiable prompts into new training questions for RL.