LongWoF-Bench: Evaluating EvoMap Genes for Verifiable Long-Workflow Tasks

Hugging Face Daily Papers Papers

Summary

The paper presents LongWoF-Bench, a benchmark for evaluating long-workflow tasks, and demonstrates that EvoMap Gene enhances task completion efficiency and reduces token costs by reusing verified execution experience.

Large language models are increasingly expected to execute complex workflows whose success depends on maintaining interdependent constraints and producing artifacts that satisfy strict end-to-end verification. Yet successful execution experience is typically lost after a single run, forcing subsequent models to rediscover strategies and failure modes from scratch. We study whether such experience can instead be externalized and reused through EvoMap, where verifier-confirmed execution trajectories are consolidated into structured Gene. To evaluate this setting, we introduce the Long-Workflow Benchmark (LongWoF-Bench), comprising 778 machine-verifiable tasks across code generation, agent-environment synthesis, mathematical reasoning, and rule following. On the 252 tasks with verifier-confirmed Opus trajectories, evolved EvoMap Gene outperform Skill across all seven evaluated models by 8.7-15.5 percentage points, with the gains extending to consumer models from different model families. In contrast, reference-distilled Gene do not exhibit the same advantage, indicating that compact representation alone is insufficient and that Gene utility is closely associated with verified experience provenance. For Claude Opus, Gene reuse also completes 39 more tasks than Skill while reducing solve-time token consumption by 9.9%. Together, these results show that verified execution experience can be retained and shared as a reusable external resource, enabling models to improve long-workflow completion without repeatedly paying the full cost of experience discovery.
Original Article
View Cached Full Text

Cached at: 08/25/26, 12:35 PM

Paper page - LongWoF-Bench: Evaluating EvoMap Genes for Verifiable Long-Workflow Tasks

Source: https://huggingface.co/papers/2608.23200

Abstract

EvoMap externalizes verified execution experience into reusable structured Gene, improving long-workflow task completion and reducing token costs across diverse models.

Large language models are increasingly expected to execute complex workflows whose success depends on maintaining interdependent constraints and producing artifacts that satisfy strict end-to-end verification. Yet successful execution experience is typically lost after a single run, forcing subsequent models to rediscover strategies and failure modes from scratch. We study whether such experience can instead be externalized and reused throughEvoMap, whereverifier-confirmed execution trajectoriesare consolidated into structuredGene. To evaluate this setting, we introduce the Long-Workflow Benchmark (LongWoF-Bench), comprising 778 machine-verifiable tasks across codegeneration, agent-environment synthesis, mathematical reasoning, and rule following. On the 252 tasks with verifier-confirmed Opus trajectories, evolvedEvoMapGeneoutperformSkillacross all seven evaluated models by 8.7-15.5 percentage points, with the gains extending to consumer models from different model families. In contrast,reference-distilled Genedo not exhibit the same advantage, indicating that compact representation alone is insufficient and thatGeneutility is closely associated withverified experience provenance. For Claude Opus,Genereuse also completes 39 more tasks thanSkillwhile reducing solve-time token consumption by 9.9%. Together, these results show that verified execution experience can be retained and shared as a reusable external resource, enabling models to improve long-workflow completion without repeatedly paying the full cost of experience discovery.

View arXiv pageView PDFAdd to collection

Get this paper in your agent:

hf papers read 2608\.23200

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2608.23200 in a model README.md to link it from this page.

Datasets citing this paper1

#### EvoMapAI/LongWoF-Bench Viewer• Updated38 minutes ago • 1.74k • 69

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2608.23200 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles

ContextWeave: A Real-World Workflow Benchmark

arXiv cs.AI

ContextWeave is a new longitudinal benchmark that evaluates whether recalled memory improves downstream agent performance in realistic office-work streams, using privacy-preserved multi-month workflows of 14 participants to create 1,005 executable tasks.

EvoMAS: Learning Execution-Time Workflows for Multi-Agent Systems

arXiv cs.AI

EvoMAS is a framework for learning execution-time workflows in multi-agent systems by formulating workflow construction as a sequential decision problem. It outperforms static multi-agent design methods on complex tasks by adapting agent coordination dynamically based on evolving task states.