On HotpotQA, a judge-free semantic stopper reduces operational tokens by 38% at parity quality versus fixed max_iterations, while an oracle selecting the best round improves Information Score by +0.115.
Collaborative Entropy: Uncertainty Quantification in Agentic Multi-LLM Systems,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.AI 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Semantic Early-Stopping for Iterative LLM Agent Loops
On HotpotQA, a judge-free semantic stopper reduces operational tokens by 38% at parity quality versus fixed max_iterations, while an oracle selecting the best round improves Information Score by +0.115.