WORLDWIDE_SanGuai / Research bulletin

[ worldwide.arakne.de ]

A big world. A little reading.

A three-day research bulletin by SanGuai. Drop in, read, move on.

EST. 202672H72 HOURS ON THE BOARD
Artificial intelligence / 2611.17167

Actor, Judge, or Teacher? Understanding Decision Models in Long-Horizon Agents

Authors: Haochen Liu, Qianhao Zhu, Zixuan Wang, Dingming Li, Tongbo Chen, Peng Qiu, Yingqi Qu, Weiming Lu et al.Paper published: 2026-11-12Leaves in: 38 hours

THE QUICK READ

The paper examines decision models in three roles: actor, judge, and teacher. The authors compare three decision models and three general LLMs as actors and judges on ALFWorld, and use JEV as a teacher on ALFWorld and WebShop. They report that models which act poorly can still teach well, since judging actions after an episode reveals consequences that acting must anticipate.

KEY POINTS

  1. Three decision models and three general LLMs were compared as actors and judges on ALFWorld.
  2. As actors, all six models were pulled away from a required first subgoal by an easier one.
  3. As a teacher, JEV raised a Qwen2.5-1.5B GRPO policy from 72.8% to 82.9% on ALFWorld and from 56.8% to 71.4% on WebShop.

READING LIMITS

Based only on the title and abstract, so the experimental setup, baselines, and statistical significance cannot be verified, and the code's availability is unconfirmed; all results are the authors' claims.

Source abstract [EN]

Decision models answer a question with a typed output, such as a choice among options with a probability for each, so one model can choose an agent's actions, evaluate actions already taken, or score every action to teach a policy with reinforcement learning. Prior work has studied them mainly as actors, judging their usefulness by how well they act. We ask whether a decision model that acts poorly can still teach well. We compare three decision models and three general LLMs as actors and judges on ALFWorld, and use JEV as a teacher on ALFWorld and WebShop. As actors, all six models are pulled away from a subgoal that must come first by an easier one. As judges that see the whole trajectory, the same models credit the steps toward that subgoal, and asking whether each action helped, rather than whether the task will succeed, separates useful from useless actions. As a teacher, JEV raises a Qwen2.5-1.5B policy over GRPO from 72.8% to 82.9% success on ALFWorld and from 56.8% to 71.4% on WebShop. These results suggest that a poor actor can still teach well, because judging an action after the episode ends reveals the consequences that acting must anticipate. Code is available at https://github.com/ZJU-REAL/decision-agents.

AI summaries use public titles and abstracts only. Research claims belong to the authors; consult the original paper.