trajectory-judge: What Outcome-Only LLM Judges Miss on Agent Trajectories

• URL: http://arxiv.org/abs/2609.00038v1
• Date: Sat, 29 Aug 2026 10:14:05 GMT
• Title: trajectory-judge: What Outcome-Only LLM Judges Miss on Agent Trajectories
• Abstract summary: Outcome-only evaluation is the production default for LLM agents.<n>We measure that blind spot where ground truth is known by construction.<n>We argue that judge evaluations must stratify recall by outcome survival.
• Score: 0.0
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Outcome-only evaluation is the production default for LLM agents: show a judge the request and the final reply and ask whether it was handled well. The metric is structurally blind to an agent that reaches the right answer the wrong way. We measure that blind spot where ground truth is known by construction: a deterministic tool-using support-desk environment, a scripted oracle policy that always solves it, and a fault injector that breaks exactly one thing at a known step, stratifying faults by whether the customer-visible outcome survived (silent) or not (loud). Five judges (programmatic rules, outcome-only, step-rubric at two model sizes, and a self-consistency ensemble) are scored on detection, step localisation, fault typing, calibration, and cost over 400 trajectories. The outcome-only judge catches 84% of loud faults but 45% of silent ones while flagging 33% of correct trajectories; a step-rubric judge reaches 77% silent recall with zero false alarms at 3x the cost. No judge reads the final reply: an invented promise appended to an otherwise perfect trajectory evades the rules entirely and the step judge 82% of the time, and self-consistency triples cost while improving nothing. We argue that judge evaluations must stratify recall by outcome survival, and release the environment, the injector, all raw verdicts, and an analysis pipeline that rebuilds every number offline.

Related papers

• LLM-as-a-Judge Is Not an Oracle: Why Self-Improving Agents Need Deterministic Guardrails [0.0]
  Self-improving agent pipelines have a problem at their center.<n>An evaluation signal failed in eleven classes: judge bias, harness and metric failures, ground-truth errors, and reward hacking.<n>We describe PROCTOR, a Teacher-Student loop in which a stateful orchestrator holds all tool access, stateless subagents diagnose failures and draft mutations they cannot apply.
  arXiv  Detail & Related papers  (2026-09-02T07:54:23Z)
• TTPO: Test-Time Policy Optimization [53.81524570216593]
  Test-Time Policy Optimization (TTPO) is an asymmetric objective that distills agreeing rollouts via OPSD and penalizes disagreeing rollouts with Grouped RL.<n>We show that TTPO matches label-supervised OPSD on five competition-level benchmarks, raises Qwen3-1.7B from 38.0% to 45.2% in TTT, yields +25.2% to +36.4% without thinking, and shows strong cross-task generalization.
  arXiv  Detail & Related papers  (2026-08-27T17:58:10Z)
• Critic Experience Bank: Self-Evolving Step-Level Confidence Estimation for LLM Agents [7.774009529494805]
  LLM agents act in external environments where each action changes the state that later decisions condition on, and where a single wrong step can waste interaction budget or trigger irreversible side effects long before the final failure is observed.<n>We introduce the method (methodshort), a self-evolving critic framework in which an LLM critic accumulates evidence from its own past judgments and their observed consequences.
  arXiv  Detail & Related papers  (2026-07-14T06:17:09Z)
• More Convincing, Not More Correct: Self-Play Reward Hacking of Reference-Free LLM Judges [4.167333498582309]
  Training a language model against its own reference-free judgments assumes a model's verdict on a shown answer tracks correctness.<n>We show it fails structurally: conditioned on a candidate, a judge scores plausibility, not correctness, leaving false-positive basins a policy learns to exploit.
  arXiv  Detail & Related papers  (2026-07-07T06:59:30Z)
• LivingArena: Do LLMs Know What Other LLMs Don't? Peer-Probing as Scalable Evaluation [79.03892269184145]
  LivingArena is an automated, contamination-resistant evaluation framework.<n>Models take turns proposing questions, aiming to pose items that opponents cannot answer correctly.<n>Questioners are encouraged to actively identify and exploit opponents' knowledge boundaries.
  arXiv  Detail & Related papers  (2026-06-19T06:20:58Z)
• Catching One in Five: LLM-as-Judge Blind Spots in Production Multi-Turn Transaction Agents [45.148328075418156]
  We study a deployed multi-turn food-and-beverage ordering agent and measure how many genuine quality problems it catches.<n>Our blind-spot taxonomy shows the failure is structured, not random.<n>For production multi-turn agents, automated judging is a regression floor, not a substitute for human review.
  arXiv  Detail & Related papers  (2026-06-09T02:11:01Z)
• Step-wise Rubric Rewards for LLM Reasoning [72.17879367869503]
  Reinforcement Learning with Verifiable Rewards (RLVR) is widely used to improve reasoning in large language models.<n>We find 18.2% of steps in correct responses are wrong yet positively rewarded.<n>Across six mathematical reasoning benchmarks, SRaR improves average accuracy over RaR by 3.57 points.
  arXiv  Detail & Related papers  (2026-05-17T07:08:14Z)
• A Coin Flip for Safety: LLM Judges Fail to Reliably Measure Adversarial Robustness [57.510025257780306]
  We show that existing validation protocols fail to account for substantial distribution shifts inherent to red-teaming.<n>We propose ReliableBench, a benchmark of behaviors that remain more consistently judgeable, and JudgeStressTest, a dataset designed to expose judge failures.
  arXiv  Detail & Related papers  (2026-02-04T15:13:35Z)
• Are LLM Evaluators Really Narcissists? Sanity Checking Self-Preference Evaluations [3.262230127283452]
  We show that evaluators may deliver self-preferring verdicts when the judge responds to queries which they completed incorrectly themselves.<n>We introduce an Evaluator Quality Baseline, which compares the probability that a judge incorrectly votes for itself against the probability that it votes for an incorrect response from another model.
  arXiv  Detail & Related papers  (2026-01-30T04:38:18Z)
• Evaluating Judges as Evaluators: The JETTS Benchmark of LLM-as-Judges as Test-Time Scaling Evaluators [66.83088028268318]
  This paper introduces the Judge Evaluation for Test-Time Scaling benchmark.<n>It evaluates judge performance in three domains (math reasoning, code generation, and instruction following) under three task settings.<n>Our benchmark shows that while judges are competitive with outcome reward models in reranking, they are consistently worse than process reward models in beam search procedures.
  arXiv  Detail & Related papers  (2025-04-21T17:33:23Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.