AIHugging Face
Heuristic editor, no API keyVerdict: NotableTraceDance: An Automated System for Building Agent Behavior Benchmarks from Real-World Agent Deployment Traces
An agent can complete a task while exhibiting undesirable behavior during execution.
Key numbers
- 95.3% of build-target requests
- 84% of sampled instances
VerdictWorth a reader's time today.
Abstract
An agent can complete a task while exhibiting undesirable behavior during execution. Developers need tests for the specific behaviors encountered in deployment, beyond fixed benchmark suites. We present TraceDance, an agent system that constructs targeted benchmarks from deployment traces for user-specified undesirable behaviors. For efficient construction, Anchor-and-Confirm combines programmable retrieval with candidate-level confirmation by a Flash large language model (LLM), while the Anchor Synthesis Loop generates and revises specifications for custom behaviors. The benchmarks use decision-point continuation to evaluate an LLM's next turn at a recorded decision point with a behavior-specific rubric, without a reference answer or environment replay. Experiments in coding and general tool use draw on 252,557 sessions and produce 107 benchmarks with 4,125 instances, fulfilling 95.3% of build-target requests. Both human annotators confirm the requested behavior in 84% of sampled instances, and the automated grader's agreement with human pass/fail judgments is comparable to that between the annotators. Nine frontier LLMs achieve a mean pass rate of only 26.7%, showing that they still struggle to respond appropriately at the evaluated decision points. Analysis across behavior-specific benchmarks further reveals weaknesses in how current LLMs behave as agents. By turning deployment problems into targeted benchmarks, TraceDance could serve as a key component of the recursive self-improvement (RSI) loop.
The editor's rubric
| Dimension | Level | Weight | What that level means |
|---|---|---|---|
| Leverage | ████░ 4 | 24% | A general-purpose tool used across several fields (Adam, ResNet, LoRA, next-generation sequencing). |
| Magnitude | ██░░░ 2 | 18% | Solid incremental gain on a meaningful problem. |
| Evidence | ███░░ 3 | 14% | Solid: multiple benchmarks or cohorts, ablations, fair baselines, released code or data. |
| Novelty | ██░░░ 2 | 16% | A new combination of known ideas. |
| Trajectory | ███░░ 3 | 18% | A clear path to scale. |
| Stakes | ██░░░ 2 | 10% | Benefits a professional community (practitioners, clinicians, engineers). |
Editor’s rationale
Heuristic triage from title and abstract text only, not a reading of the paper. Cues found: method (we propose); breadth (many tasks, programmable); verification (multiple benchmarks); scale (efficient); stakes (general AI).
How the score was computed
- Merit
- 5.6 / 10
- Adjusted merit
- 4.7 / 10
- Attention
- 83%
- Freshness
- 32%
- Citations0 (reference 15, via semantic-scholar, Sep 29, 2026, 23:53 UTC)
- Influential citations0 (reference 3, via semantic-scholar, Sep 29, 2026, 23:53 UTC)
- Hugging Face upvotes62 (reference 25, via hf-daily, Oct 1, 2026, 02:16 UTC)
- GitHub stars9 (reference 250, via hf-daily, Oct 1, 2026, 02:16 UTC)