What Makes Interaction Trajectories Effective for Training Terminal Agents?

• URL: http://arxiv.org/abs/2606.03461v1
• Date: Tue, 02 Jun 2026 10:37:47 GMT
• Title: What Makes Interaction Trajectories Effective for Training Terminal Agents?
• Abstract summary: Terminal-Lego is a scalable pipeline that transforms real-world issues into environment-verified agentic tasks.<n>Students fine-tuned on trajectories from DeepSeek-V3.2, a lower-scoring agent, exhibit significantly stronger generalization.<n>With only 15.3k Terminal-Lego trajectories, Qwen3-32B achieves a 24.3% score on Terminal-Bench 2.0, rivaling previous SOTA performance established with over 30x the data volume.
• Score: 55.62817294510983
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Stronger code agents are commonly assumed to be superior teachers for post-training, yet this assumption remains poorly disentangled from task difficulty, harness design, and student capacity. We investigate this pedagogical link using Terminal-Lego, a scalable pipeline that transforms multi-domain real-world issues into environment-verified agentic tasks. Surprisingly, standalone performance does not dictate teaching efficacy: while Claude Opus 4.6 achieves higher scores on Terminal-Bench 2.0, students fine-tuned on trajectories from DeepSeek-V3.2, a lower-scoring agent, exhibit significantly stronger generalization. We attribute this "pedagogical paradox" to Environment-Grounded Supervision (EGS): trajectories that explicitly expose inspect-act-verify behaviors through harness-visible interactions allow students to internalize robust problem-solving routines rather than fragile action sequences. Scaling analysis reveals exceptional data efficiency: with only 15.3k Terminal-Lego trajectories, for example, Qwen3-32B achieves a 24.3% score on Terminal-Bench 2.0, rivaling previous SOTA performance established with over 30x the data volume. Our results suggest that the frontier of agent post-training lies beyond mere outcome-matching, shifting the focus toward "Harness Engineering", where the systematic design of environment-grounded interaction structures serves as the primary catalyst for reproducible and generalizable agentic intelligence.

Related papers

• Cross-Benchmark Generalization in Long-Horizon Agents [1.0742728354283815]
  We post-train an open-weight mixture-of-experts model on 363 long-horizon Model Context Protocol (MCP) tasks across 27 categories.<n>No externalbenchmark task or grader entered training and no external score informed the reward, training hyper parameters, trained-checkpoint selection, or stopping.<n>Both software benchmarks improve despite the training collection containing no software-engineering tasks.
  arXiv  Detail & Related papers  (2026-07-31T18:05:36Z)
• Filtering Harmful Actions Isn't Enough: Phantom Transfer in Agentic SDF [0.0]
  We investigate the effects of training on synthetic agentic trajectories containing adversarial interactions.<n>Finetuning on these trajectories consistently increases misaligned behavior.<n>Our results suggest that action level filtering is insufficient to ensure the safety of synthetic agentic training data.
  arXiv  Detail & Related papers  (2026-07-12T13:00:41Z)
• Terminal-World: Scaling Terminal-Agent Environments via Agent Skills [52.39713754337834]
  We introduce Terminal-World, a fully automated pipeline that uses agent skills as the central synthesis primitive.<n>We construct 5,723 training environments and train Terminal-World-8B/14B/32B, evaluated across 6 benchmarks.<n>Terminal-World-32B surpasses Nemotron-Terminal-32B on Terminal-Bench 2.0 by +4.5 Pass@1 (31.5) and achieves 43.8 Pass@3.
  arXiv  Detail & Related papers  (2026-05-20T08:14:51Z)
• TCOD: Exploring Temporal Curriculum in On-Policy Distillation for Multi-turn Autonomous Agents [55.27396165691312]
  We identify a key limitation of vanilla OPD in multi-turn agent settings, which we term Trajectory-Level KL Instability.<n>We propose TCOD, a framework that controls the trajectory depth exposed to the student and progressively expands it with a curriculum schedule.<n> Experimental results across four student-teacher pairs show that TCOD mitigates KL escalation and enhances KL stability throughout training, improving agent performance by up to 18 points over vanilla OPD.
  arXiv  Detail & Related papers  (2026-04-27T03:38:27Z)
• Structured Distillation of Web Agent Capabilities Enables Generalization [33.41723014134473]
  We introduce Agent-as-Annotators, a framework that structures synthetic trajectory generation for web agents by analogy to human annotation roles.<n>Using Gemini 3 Pro as teacher, we generate 3,000 trajectories across six web environments.<n>The resulting model achieves 41.5% on WebArena, surpassing closed-source models such as Claude 3.5 Sonnet (36.0%) and GPT-4o (31.5%) under the same evaluation protocol.
  arXiv  Detail & Related papers  (2026-04-09T04:04:15Z)
• Claw-Eval: Toward Trustworthy Evaluation of Autonomous Agents [66.97968363332465]
  We introduce Claw-Eval, an end-to-end evaluation suite addressing all three gaps in agent benchmarks.<n>Claw-Eval comprises 300 human-verified tasks spanning 9 categories across three groups.<n>Every agent action is recorded through three independent evidence channels.
  arXiv  Detail & Related papers  (2026-04-07T17:43:18Z)
• TopoCurate:Modeling Interaction Topology for Tool-Use Agent Training [53.93696896939915]
  Training tool-use agents typically rely on Supervised Fine-Tuning (SFT) on successful trajectories and Reinforcement Learning (RL) on pass-rate-selected tasks.<n>We propose TopoCurate, an interaction-aware framework that projects multi-trial rollouts from the same task into a unified semantic quotient topology.<n>TopoCurate achieves consistent gains of 4.2% (SFT) and 6.9% (RL) over state-of-the-art baselines.
  arXiv  Detail & Related papers  (2026-03-02T10:38:54Z)
• Eigen-1: Adaptive Multi-Agent Refinement with Monitor-Based RAG for Scientific Reasoning [53.45095336430027]
  We develop a unified framework that combines implicit retrieval and structured collaboration.<n>On Humanity's Last Exam (HLE) Bio/Chem Gold, our framework achieves 48.3% accuracy.<n>Results on SuperGPQA and TRQA confirm robustness across domains.
  arXiv  Detail & Related papers  (2025-09-25T14:05:55Z)
• AWorld: Orchestrating the Training Recipe for Agentic AI [35.94278765364194]
  We introduce AWorld, an open-source system engineered for large-scale agent-environment interaction.<n>By distributing tasks across a cluster, AWorld accelerates experience collection by 14.6x compared to standard single-node, sequential execution.<n>We trained a Qwen3-32B-based agent that achieves pass@1 accuracy of 32.23% on the GAIA test set.
  arXiv  Detail & Related papers  (2025-08-28T04:04:30Z)
• Training Long-Context, Multi-Turn Software Engineering Agents with Reinforcement Learning [29.605396813225386]
  We show how reinforcement learning can be used to train agents for multi-turn interactive tasks.<n>Our methodology offers a practical approach for training capable agents for multi-turn interactive tasks using open-weight models.
  arXiv  Detail & Related papers  (2025-08-05T14:30:47Z)

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.