Search papers, labs, and topics across Lattice.
This paper investigates whether Large Language Models (LLMs) can effectively adapt to and act upon unknown communication conventions in a cooperative setting, using a Hanabi-like game with scripted hints. While LLMs can decode sender intents with reasonable accuracy, their subsequent actions do not consistently align with these decoded conventions, indicating a gap between knowledge representation and behavioral execution. The study finds that providing explicit action recommendations, rather than general rule statements, significantly improves cooperative performance, suggesting a need for more direct guidance to translate inferred knowledge into effective actions.
LLMs can decode partner intents but struggle to translate this knowledge into cooperative actions, with explicit action recommendations proving far more effective than abstract rule statements.
Cooperation with unfamiliar partners requires adapting to communication conventions that are not known in advance. We study this problem in a controlled Hanabi-derived environment with scripted hint generation, LLM-controlled receiving decisions, and frozen model weights. Across eight LLMs, linear probes recover intent conventions substantially more accurately than target conventions, yet receiving choices do not consistently agree with the sender's convention. We compare probe-predicted and ground-truth conventions presented either as general rules or as externally computed action recommendations. Rule statements yield modest and model-dependent changes in cooperation, whereas action translation produces larger gains on average. In a Qwen3-8B case study, matched-state statement reversals reveal much greater sensitivity to action recommendations than to rule statements. Activation transfers from oracle-action and non-oracle hint-restatement donors improve intent accuracy on both action classes, but the tested alternatives do not reliably reproduce these benefits. Together, these results distinguish convention decodability, sensitivity to convention information, and cooperative performance, and highlight limitations in turning available partner information into receiving decisions.
