[Re] Benchmarking LLM Capabilities in Negotiation through Scoreable Games

TL;DR

This paper critically evaluates a complex negotiation benchmark for Large Language Models, replicating experiments, introducing new metrics, and analyzing model behaviors to assess its robustness, objectivity, and usability in multi-agent negotiation tasks.

Contribution

It reproduces and extends the original benchmark, identifies limitations, and provides insights into its applicability and the importance of context in model evaluation.

Findings

Benchmark is complex but comparison is ambiguous

Limitations in information leakage detection

Context significantly impacts model evaluation

Abstract

Large Language Models (LLMs) demonstrate significant potential in multi-agent negotiation tasks, yet evaluation in this domain remains challenging due to a lack of robust and generalizable benchmarks. Abdelnabi et al. (2024) introduce a negotiation benchmark based on Scoreable Games, with the aim of developing a highly complex and realistic evaluation framework for LLMs. Our work investigates the reproducibility of claims in their benchmark, and provides a deeper understanding of its usability and generalizability. We replicate the original experiments on additional models, and introduce additional metrics to verify negotiation quality and evenness of evaluation. Our findings reveal that while the benchmark is indeed complex, model comparison is ambiguous, raising questions about its objectivity. Furthermore, we identify limitations in the experimental setup, particularly in information leakage detection and thoroughness of the ablation study. By examining and analyzing the behavior of a wider range of models on an extended version of the benchmark, we reveal insights that provide additional context to potential users. Our results highlight the importance of context in model-comparative evaluations.