MERIT Feedback Elicits Better Bargaining in LLM Negotiators

Researchers introduce AgoraBench, a new framework for improving Large Language Models' bargaining and negotiation capabilities through utility-based feedback mechanisms. The study reveals that current LLMs struggle with strategic depth in negotiations and proposes human-aligned metrics and training methods to enhance their performance.

• →AgoraBench benchmark covers nine challenging negotiation scenarios including deception and monopoly situations.
• →Current LLMs demonstrate limited strategic depth and struggle to adapt to complex human factors in bargaining.
• →The framework introduces human-aligned metrics based on utility theory including agent utility, negotiation power, and acquisition ratio.
• →Baseline LLM negotiation strategies often diverge significantly from human preferences and expectations.
• →The proposed training pipeline using prompting and finetuning substantially improves LLM negotiation performance and strategic awareness.