A new escape-room benchmark measures creative reasoning in language-model agents, and the EscapeAgent framework (Foresight plus Reflection) reduces hints and steps by up to 40%.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
EscapeBench: Towards Advancing Creative Intelligence of Language Model Agents
A new escape-room benchmark measures creative reasoning in language-model agents, and the EscapeAgent framework (Foresight plus Reflection) reduces hints and steps by up to 40%.