Your Policy Is Secretly a Harness: The Decision Geometry of Agent Interfaces
The paper argues that language model agents are trained and deployed through a harness of prompts, tools, and control logic, and that a trained policy absorbs preferences from its harness. Under KL-regularized control, deployed return is decomposed into the harness's ceiling minus the policy's mismatch. Training reduces only mismatch, while harness changes move both terms, so fixed-model comparisons conflate the two. Experiments on MiniWoB++ and Terminal-Bench Lite support these claims, including a stronger-model stand-in that reproduces a coding-agent accuracy ranking.
Open original source huggingface.coOne primary source; other links are context, not independent corroboration.
- Primary sourceHugging Face Daily Papers
- Discovery linkHugging Face
- Linked in sourceView arXiv page
- Linked in sourceView PDF
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 argues that language model agents are trained and deployed through a harness of prompts, tools, and control logic, and that a trained policy absorbs preferences from its harness. Under KL-regularized control, deployed return is decomposed into the harness's ceiling minus the policy's mismatch. Training reduces only mismatch, while harness changes move both terms, so fixed-model comparisons conflate the two. Experiments on MiniWoB++ and Terminal-Bench Lite support these claims, including a stronger-model stand-in that reproduces a coding-agent accuracy ranking.
A READER'S EVIDENCE CHECKHow to read this signalSix questions about claims, sources, risks, and open checks
This is a signal-level guide until claim-level evidence is available. “Not recorded” means the evidence has not been assessed here; it does not mean a result is false. How we check AI news →
- 01Who is making each claim?
The claimant and whether each claim was measured or asserted are not recorded in this public brief.
- 02What was actually tested?
Not recorded per claim. Check the source for sample size, tasks, model and benchmark versions, and reproducibility.
- 03Where are the sources?
Original link: huggingface.co. Brief basis: fetched source page. Primary page date and quote locations are not established in this panel.
- 04What could mislead?
Source claims not independently checked; results are reported on specific benchmarks (MiniWoB++, Terminal-Bench Lite) and a stronger model is used as a stand-in for the optimum.
- 05What is at risk if wrong?
Not assessed per claim. Consider safety, privacy, cost, and adoption decisions before relying on the result.
- 06What would change the view?
No claim-specific check is recorded. An inspectable primary method or independent replication could change the assessment.
Key findings from the stored brief
- Under KL-regularized control, the log odds of an optimal policy recover how strongly its harness favors each action.
- Deployed return equals harness ceiling minus policy mismatch; training changes only mismatch.
- Reversing observation field order left optimal decisions unchanged but lowered a trained model's MiniWoB++ success from 85.0% to 19.2%.
Source claims not independently checked; results are reported on specific benchmarks (MiniWoB++, Terminal-Bench Lite) and a stronger model is used as a stand-in for the optimum.
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.
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.
A publication revision history is not stored for this signal. New engagement snapshots do not mean the source itself changed.
Why this may matter
Harnesses are often judged by how an already trained model performs under them. The paper suggests such comparisons mix what a harness offers with how well the model uses it, which could mislead agent setup decisions.
Limits and open questions
Source claims not independently checked; results are reported on specific benchmarks (MiniWoB++, Terminal-Bench Lite) and a stronger model is used as a stand-in for the optimum.
Organization-specific financial, security, and operational effects require the organization's own context and review.
Discussion
Comments are reviewed before publication. Discuss the evidence, not the person.
No published comments yet.
Reader discussion opens with accounts. You can read the original source and published comments now.