Estimating the Empowerment of Language Model Agents

2025-09-26

AI Summary

Existing evaluation of language model (LM) agents relies heavily on manually designed tasks, suffering from poor scalability and limited generalizability. Method: This paper introduces an information-theoretic empowerment-based evaluation framework—first applying the concept of empowerment (defined as mutual information between actions and future states) to LM agent assessment—and proposes the EELMA algorithm, which estimates effective empowerment via unsupervised, multi-turn textual interaction. The approach requires no task-specific annotations and operates in open-ended environments such as language games and web browsing. Contribution/Results: Experiments demonstrate strong correlation between empowerment and downstream task performance; reveal significant influences of environmental complexity, chain-of-thought capability, model scale, and memory length on empowerment; and enable identification of high-empowerment critical states. This work establishes a theoretically grounded, empirically practical paradigm for scalable, task-agnostic LM agent evaluation.

Technology Category

• Multiagent Systems: Multiagent Learning
• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: (Large) Language Models

Application Category

• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

As language model (LM) agents become more capable and gain broader access to real-world tools, there is a growing need for scalable evaluation frameworks of agentic capability. However, conventional benchmark-centric evaluations are costly to design and require human designers to come up with valid tasks that translate into insights about general model capabilities. In this work, we propose information-theoretic evaluation based on empowerment, the mutual information between an agent's actions and future states, as an open-ended method for evaluating LM agents. We introduce EELMA (Estimating Empowerment of Language Model Agents), an algorithm for approximating effective empowerment from multi-turn text interactions. We validate EELMA on both language games and scaled-up realistic web-browsing scenarios. We find that empowerment strongly correlates with average task performance, characterize the impact of environmental complexity and agentic factors such as chain-of-thought, model scale, and memory length on estimated empowerment, and that high empowerment states and actions are often pivotal moments for general capabilities. Together, these results demonstrate empowerment as an appealing general-purpose metric for evaluating and monitoring LM agents in complex, open-ended settings.

Problem

• Evaluating language model agents' capabilities without predefined benchmarks
• Measuring agent empowerment through action-state mutual information
• Validating empowerment correlation with performance in complex environments

Innovation

• Information-theoretic evaluation using empowerment metric
• EELMA algorithm for estimating empowerment from text
• Validated on language games and web-browsing scenarios