CLARITI uses Shapley-based empirical rewards in RL to ask fewer but more effective clarifying questions in SE tasks, matching GPT-5 performance with 41% fewer questions.
Does replacing string x-axis with numeric (e.g., [0,1]) or datetime reproduce? Tied to string labels or stacking/axis in- teraction?[R, repro]
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.SE 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Asking What Matters: Reward-Driven Clarification for Software Engineering Tasks
CLARITI uses Shapley-based empirical rewards in RL to ask fewer but more effective clarifying questions in SE tasks, matching GPT-5 performance with 41% fewer questions.