paper-with-me

홈 › Papers

Estimating the Empowerment of Language Model Agents

2025-09-26 · Jinyeop Song, Jeff Gore, Max Kleiman-Weiner arxiv

As language model (LM) agents become increasingly capable and adopted in real-world applications, there is a growing need for scalable evaluation frameworks beyond costly, manually designed benchmarks. We propose information-theoretic evaluation based on empowerment, an information-theoretic measure of an agent's influence on future states through its actions. To handle the unique challenges of text-based environments, we introduce EELMA (Estimating Empowerment of Language Model Agents), an algorithm for approximating effective empowerment from multi-turn text interactions. We demonstrate EELMA on textual games and realistic web and tool-use environments, showing that empowerment strongly correlates with average task performance. We further analyze how empowerment varies across models, environment complexity, and agent configurations, and show that high-empowerment states and actions often mark pivotal moments for general capabilities. These results establish empowerment as a goal-agnostic metric that complements task-success measures for LM-agent evaluation.

📄 PDF Abstract BibTeX arXiv:2509.22504

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

AvE: Assistance via Empowerment

2020-06-26 · NeurIPS 2020 12 · Yuqing Du, Stas Tiomkin, Emre Kiciman, Daniel Polani 외

One difficulty in using artificial agents for human-assistive applications lies in the challenge of accurately assisting with a person's goal(s). Existing methods tend to rely on inferring the human's goal, which is chal…

Training LLM Agents to Empower Humans

2025-10-15 · Evan Ellis, Vivek Myers, Jens Tuyls, Sergey Levine 외 arxiv

Assistive agents should not only take actions on behalf of a human, but also step out of the way and cede control when there are important decisions to be made. However, current methods for building assistive agents, whe…

Latent-Predictive Empowerment: Measuring Empowerment without a Simulator

2024-10-15 · Andrew Levy, Alessandro Allievi, George Konidaris

Empowerment has the potential to help agents learn large skillsets, but is not yet a scalable solution for training general-purpose agents. Recent empowerment methods learn diverse skillsets by maximizing the mutual info…

Reliably Re-Acting to Partner's Actions with the Social Intrinsic Motivation of Transfer Empowerment

2022-03-07 · Tessa van der Heiden, Herke van Hoof, Efstratios Gavves, Christoph Salge

We consider multi-agent reinforcement learning (MARL) for cooperative communication and coordination tasks. MARL agents can be brittle because they can overfit their training partners' policies. This overfitting can prod…

Multi-agent Reinforcement Learningreinforcement-learningReinforcement Learning (RL)

Hierarchical Empowerment: Towards Tractable Empowerment-Based Skill Learning

2023-07-06 · Andrew Levy, Sreehari Rammohan, Alessandro Allievi, Scott Niekum 외

General purpose agents will require large repertoires of skills. Empowerment -- the maximum mutual information between skills and states -- provides a pathway for learning large collections of distinct skills, but mutual…

Hierarchical Reinforcement Learning