English static mirror for SEO/GEO · AI-assisted translation · Read Chinese original

ALIGN: Rewriting Environment Feedback Wording Boosts Qwen2.5-7B on ALFWorld from 13.4% to 31.3%

Forum topic · QianXun · 2026-08-01

Summary

ALIGN (Agent-Environment Alignment via Automated Interface Generation, arXiv:2505.21055), from ModelBest and Tsinghua NLP, shows that LLM agents often fail not because of weak models but because of ambiguous environment feedback. Simply rephrasing ALFWorld's "Nothing happens" error message into an explicit instruction raised Qwen2.5-7B's success rate from 13.4% to 31.3% with no model or environment changes. ALIGN automates this with an LLM-based interface generator that analyzes failure cases, extracts hidden constraints, and repackages feedback into static rules plus dynamic observation augmentation. Across ALFWorld, ScienceWorld, WebShop, and M3ToolEval it delivers up to +45.67% average improvement, holds across Vanilla, ReAct, Self-Consistency, Self-Refine, and Planning agents, and transfers to Qwen2.5-14B, Llama 3.1-8B, and Llama 3.3-70B. Invalid-action rates drop 65% (ALFWorld) and 49% (ScienceWorld). The post argues agent-environment interface design, not model scale, is the next optimization frontier for embodied AI and agent products.

On July 31, ModelBest and the Tsinghua NLP team published *Agent-Environment Alignment via Automated Interface Generation* (arXiv:2505.21055), offering a clean answer to the question of why AI agents underperform: the problem often lies in the wording of environment feedback, not in model capability.

Core results

  • ALFWorld: Qwen2.5-7B agent success rate 13.4% → 31.3% purely by rephrasing feedback wording
  • Up to +45.67% average across four benchmarks (ALFWorld, ScienceWorld, WebShop, M3ToolEval)
  • Consistent gains across five agent algorithms: Vanilla, ReAct, Self-Consistency, Self-Refine, Planning
  • Consistent across models: interfaces generated with Qwen2.5-7B also improve Qwen2.5-14B, Llama 3.1-8B, and Llama 3.3-70B
  • Consecutive invalid actions dropped 65% on ALFWorld and 49% on ScienceWorld
  • No agent changes, no environment changes, no RLHF — just clearer feedback wording more than doubled the success rate.

    The problem: agent-environment interface mismatch

    The paper calls this *Agent-Environment Misalignment*. In ALFWorld, an agent must "go to shelf 1" before it can "examine shelf 1"; otherwise the environment returns "Nothing happens". Agents misinterpret this as "shelf 1 is empty" rather than "you must move there first", and repeatedly hit the same wall.

    A preliminary experiment changed the error to "You need to first go to receptacle before you can examine it" — a tiny wording change that lifted Qwen2.5-7B's ALFWorld success rate by 17.9 percentage points with an unchanged 7B model and unchanged environment code.

    ALIGN automates this insight: an LLM-based Interface Generator analyzes failure cases, infers hidden constraints, and repackages environment feedback. The generated interface has two components: static information (rules, field explanations) and dynamic observation augmentation (interpreting "action had no effect" as "you need to do X first").

    Why it matters

    The past two years of embodied AI discussion has focused on model backbones — VLA, diffusion policies, world models, π0.5, Helix. ALIGN offers a different view: in many settings, model capability is not the bottleneck; agents are stuck because they cannot parse the feedback describing their current state.

    The evidence is strongly cross-cutting:

  • Cross-algorithm: all five mainstream agent paradigms improve
  • Cross-scale: 7B / 8B / 14B / 70B models all improve — this is not about small models, but about feedback ambiguity itself
  • Cross-domain: embodied (ALFWorld, ScienceWorld), web navigation (WebShop), and tool use (M3ToolEval) all improve, indicating a generic bug in agent-environment interfaces
  • Practical advice for agent builders

  • If your agent keeps failing in the same spot: before tuning the model or adding RLHF, make feedback wording explicit — turn "Nothing happens" into the missing precondition, and "Invalid action" into why it is invalid and what to do instead.
  • If you run agent evaluations: low scores may reflect unfriendly evaluation interfaces. Ablate by clarifying feedback and re-running; the gain quantifies interface pollution in your benchmark.
  • If you collect VLA training data: interface quality directly shapes trajectory quality — clearer feedback yields shorter, cleaner demonstrations and less noise.
  • If you build agent harnesses: adopt auto interface generation as a decoupled module reusable across environments.
  • If your product repeats the same mistake: check the wording of error messages in the conversation history first — it may be an interface problem, not an intelligence problem.
  • Where this line of work is heading

    ALIGN is not isolated. A parallel thread is forming: Anthropic's "Unhobbling" work (removing 80% of system prompts without coding-benchmark loss), autonomous long-horizon agent runs, and models emphasizing harness- and engineering-level capability release over parameter stacking. The common theme: model weights are no longer the core variable. Unlocking existing models from bad interfaces, harnesses, and system prompts is cheaper and more immediately effective than scaling parameters.

    One-line takeaway

    ALIGN lifted Qwen2.5-7B from 13.4% to 31.3% on ALFWorld by rewriting feedback wording — and the effect holds across algorithms, scales, and domains. The next optimization point for embodied AI and agent products lies not in the model or the environment code, but in the neglected interface between them.

    References

  • Paper arXiv:2505.21055: https://arxiv.org/abs/2505.21055
  • PDF: https://arxiv.org/pdf/2505.21055
  • HTML: https://arxiv.org/html/2505.21055
  • Review with Tables 1 & 2: https://www.themoonlight.io/review/agent-environment-alignment-via-automated-interface-generation
  • ModelBest: https://modelbest.cn

Tags

#align#llm-agents#embodied-ai#alfworld#agent-environment-alignment#qwen2-5#modelbest#agent-harness

This page is an English static mirror generated for search and AI citation. It may be a full translation or structured summary of the Chinese original. Canonical interactive discussion lives on the Chinese page: https://zhichai.net/topic/178503847