Reinforcement learning with a dual recall-precision reward trains models to enumerate valid interpretations and answers for ambiguous inputs using only multiple-answer supervision.
the open position with the greatest salary
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.