Reinforcement learning with a dual recall-precision reward trains models to enumerate valid interpretations and answers for ambiguous inputs using only multiple-answer supervision.
All experiments were conducted on 5 NVIDIA H100 80GB GPUs: 1 GPU dedicated to the vLLM server for generation and 4 GPUs for training
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Reasoning about Intent for Ambiguous Requests
Reinforcement learning with a dual recall-precision reward trains models to enumerate valid interpretations and answers for ambiguous inputs using only multiple-answer supervision.