Actor, Judge, or Teacher? Understanding Decision Models in Long-Horizon Agents
The paper studies decision models that output typed decisions and can serve as an actor (choosing actions), a judge (evaluating completed actions), or a teacher (scoring every action for step-level RL rewards).
Source-based brief · Editor review not recorded · See checks and limits ↓
One primary source; other links are context, not independent corroboration.
- Primary sourceHugging Face
- Discovery linkHugging Face
- Linked in sourceView arXiv page
- Linked in sourceView PDF
- Linked in sourcehttps://github.com/ZJU-REAL/decision-agents
The paper page is linked. Peer review and findings have not been independently checked by AINewsIQ.
Open full article view →- Original paper+45
- Fetched source page+30
- No editor link review+0
- Original non-discussion link+5
This measures the stored source trail and editorial steps. It does not certify accuracy, peer review, or independent corroboration. Popularity adds no points.
View full scoring →Read the scoring rubric →What the source says
The paper studies decision models that output typed decisions and can serve as an actor (choosing actions), a judge (evaluating completed actions), or a teacher (scoring every action for step-level RL rewards). On ALFWorld, three decision models and three general LLMs were compared; as actors all six were pulled toward an easier subgoal, but as judges seeing the full trajectory they credited steps toward the prerequisite. Using JEV as a teacher for a Qwen2.5-1.5B policy trained with GRPO raised success from 72.8% to 82.9% on ALFWorld and from 56.8% to 71.4% on WebShop.
Claim explorer
Attribution, source and conditionsSource-stated points from the stored brief. Claim-level assessments have not been recorded.
Source claims not independently checked; results are reported on two benchmarks (ALFWorld, WebShop) with one policy size.
This connects a source, an attributed statement and its qualifications. It does not establish that the claim is correct.
Questions worth asking
Open questions for further reportingWhat remains unconfirmed?
Source claims not independently checked; results are reported on two benchmarks (ALFWorld, WebShop) with one policy size.
What material was available for this brief?
The recorded brief basis is fetched source page. These points are AINewsIQ paraphrases, not quotations or independent confirmations.
As actors, all six tested models were pulled toward an easier subgoal and skipped a prerequisite subgoal on ALFWorld.
Source: Hugging Face ↗Stored source-based brief · key point 1 · AINewsIQ paraphraseAs judges seeing the full trajectory, the same models credited steps that advanced the prerequisite subgoal.
Source: Hugging Face ↗Stored source-based brief · key point 2 · AINewsIQ paraphraseUsing JEV as a teacher, GRPO-trained Qwen2.5-1.5B success rose from 72.8% to 82.9% on ALFWorld and 56.8% to 71.4% on WebShop.
Source: Hugging Face ↗Stored source-based brief · key point 3 · AINewsIQ paraphraseLimits from the stored brief
Source claims not independently checked; results are reported on two benchmarks (ALFWorld, WebShop) with one policy size.
This is a source-based abstract, not independent validation of the source's claims.
Evidence & sources
These links show where the signal came from and what the source page linked to. A linked page is context, not independent corroboration.
Related archive coverage
These relationships record a citation or topic overlap, not confirmation of the finding.
The paper page is linked. Peer review and findings have not been independently checked by AINewsIQ.
Timeline & updates
- Signal dated by source or discovery feed
This timestamp places the signal in the radar. It may be the discovery post date rather than the primary page's first publication date.
- Shared brief stored
AINewsIQ prepared one abstract from fetched source page.
Brief history · 1 retained version
These are changes to AINewsIQ’s own publication. Earlier history is unavailable unless recorded here. Complete retained versions are available to the editor.
- Version 1 · published
Source-based brief stored or updated. This does not establish that the publisher changed its page.
Automated record
Source observations · 1 recent captures
A changed capture means the extracted text differed. It can reflect page furniture or extraction changes; it does not establish a substantive correction. Source captures remain private.
Retained capture; no earlier comparison available.
New engagement snapshots do not mean the source itself changed. Source observations and AINewsIQ brief revisions are separate records.
Why this may matter
The findings suggest that a model's weakness as an actor does not preclude its usefulness as a source of step-level reward for training other agents, which could change how decision models are evaluated and used in RL pipelines for long-horizon tasks.
Limits and open questions
Source claims not independently checked; results are reported on two benchmarks (ALFWorld, WebShop) with one policy size.
Organization-specific financial, security, and operational effects require the organization's own context and review.

Join the discussion
Have a different perspective or a question about the evidence? Discussion is free with a verified email account. Every comment is reviewed before publication.
No published comments yet.
New commenters: 2 submissions per 24 hours, at least 5 minutes apart, until 3 are approved. Discussion guidelines and moderation →