LLM agents achieve Nash-stable coalitions 73% of the time under a new Coalition-of-Thought protocol versus 58% and 42% with chain-of-thought and standard prompting, supported by existence proofs and bounded-rationality analysis in a hedonic game framework.
Answer the following question. Provide only the final answer
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.GT 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Coalition Formation in LLM Agent Networks: Stability Analysis and Convergence Guarantees
LLM agents achieve Nash-stable coalitions 73% of the time under a new Coalition-of-Thought protocol versus 58% and 42% with chain-of-thought and standard prompting, supported by existence proofs and bounded-rationality analysis in a hedonic game framework.