Efficient Management of LLM-Based Coaching Agents' Reasoning While Maintaining Interaction Quality and Speed

要旨

LLM-based agents improve upon standalone LLMs, which are optimized for immediate intent-satisfaction, by allowing the pursuit of more extended objectives, such as helping users over the long term. To do so, LLM-based agents need to reason before responding. For complex tasks like personalized coaching, this reasoning can be informed by adding relevant information at key moments, shifting it in the desired direction. However, the pursuit of objectives beyond interaction quality may compromise this very quality. Moreover, as the depth and informativeness of reasoning increase, so do the number of tokens required, leading to higher latency and cost. This study investigates how an LLM-based coaching agent can adjust its reasoning depth using a discrepancy mechanism that signals how much reasoning effort to allocate based on how well the objective is being met. Our discrepancy-based mechanism constrains reasoning to better align with alternative objectives, reducing cost roughly tenfold while minimally impacting interaction quality.

著者
Andreas Göldi
University of St.Gallen, St.Gallen, Switzerland
Roman Rietsche
Institute for Digital Technology Management, Bern, Bern, Switzerland
Lyle Ungar
University of Pennsylvania, Philadelphia, Pennsylvania, United States
DOI

10.1145/3706598.3713606

論文URL

https://dl.acm.org/doi/10.1145/3706598.3713606

動画

会議: CHI 2025

The ACM CHI Conference on Human Factors in Computing Systems (https://chi2025.acm.org/)

セッション: Sports

G304
6 件の発表
2025-04-30 01:20:00
2025-04-30 02:50:00
日本語まとめ
読み込み中…