Saved in:
Bibliographic Details
Main Authors: Costarelli, Anthony, Allen, Mat, Hauksson, Roman, Sodunke, Grace, Hariharan, Suhas, Cheng, Carlson, Li, Wenjie, Clymer, Joshua, Yadav, Arjun
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2406.06613
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914881156939776
author Costarelli, Anthony
Allen, Mat
Hauksson, Roman
Sodunke, Grace
Hariharan, Suhas
Cheng, Carlson
Li, Wenjie
Clymer, Joshua
Yadav, Arjun
author_facet Costarelli, Anthony
Allen, Mat
Hauksson, Roman
Sodunke, Grace
Hariharan, Suhas
Cheng, Carlson
Li, Wenjie
Clymer, Joshua
Yadav, Arjun
contents Large language models have demonstrated remarkable few-shot performance on many natural language understanding tasks. Despite several demonstrations of using large language models in complex, strategic scenarios, there lacks a comprehensive framework for evaluating agents' performance across various types of reasoning found in games. To address this gap, we introduce GameBench, a cross-domain benchmark for evaluating strategic reasoning abilities of LLM agents. We focus on 9 different game environments, where each covers at least one axis of key reasoning skill identified in strategy games, and select games for which strategy explanations are unlikely to form a significant portion of models' pretraining corpuses. Our evaluations use GPT-3 and GPT-4 in their base form along with two scaffolding frameworks designed to enhance strategic reasoning ability: Chain-of-Thought (CoT) prompting and Reasoning Via Planning (RAP). Our results show that none of the tested models match human performance, and at worst GPT-4 performs worse than random action. CoT and RAP both improve scores but not comparable to human levels.
format Preprint
id arxiv_https___arxiv_org_abs_2406_06613
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle GameBench: Evaluating Strategic Reasoning Abilities of LLM Agents
Costarelli, Anthony
Allen, Mat
Hauksson, Roman
Sodunke, Grace
Hariharan, Suhas
Cheng, Carlson
Li, Wenjie
Clymer, Joshua
Yadav, Arjun
Computation and Language
Artificial Intelligence
Large language models have demonstrated remarkable few-shot performance on many natural language understanding tasks. Despite several demonstrations of using large language models in complex, strategic scenarios, there lacks a comprehensive framework for evaluating agents' performance across various types of reasoning found in games. To address this gap, we introduce GameBench, a cross-domain benchmark for evaluating strategic reasoning abilities of LLM agents. We focus on 9 different game environments, where each covers at least one axis of key reasoning skill identified in strategy games, and select games for which strategy explanations are unlikely to form a significant portion of models' pretraining corpuses. Our evaluations use GPT-3 and GPT-4 in their base form along with two scaffolding frameworks designed to enhance strategic reasoning ability: Chain-of-Thought (CoT) prompting and Reasoning Via Planning (RAP). Our results show that none of the tested models match human performance, and at worst GPT-4 performs worse than random action. CoT and RAP both improve scores but not comparable to human levels.
title GameBench: Evaluating Strategic Reasoning Abilities of LLM Agents
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2406.06613