Paper page - What Else Needs Fixing? Exploring Cost-Effective Test-Time Compute for Revision Propagation in Artifacts Generated Through Conversation · AuraTracer

SummaryAI generated

2026 年 9 月 8 日，Hugging Face Papers 发布论文《What Else Needs Fixing? Exploring Cost-Effective Test-Time Compute for Revision Propagation in Artifacts Generated Through Conversation》，探讨通过对话生成的艺术品中修订传播所需的低成本测试时间计算。该研究旨在优化相关流程效率。

Related eventsRELATED EVENTS

2026-09-09 05:29OpenAI claims huge maths breakthrough on a famed ‘Millennium Problem’
2026-09-09 06:11OpenAI claims to have solved the Navier-Stokes equation problem, but its originality is questioned by the academic community
2026-09-08 22:00Google DeepMind Maps 9 Billion Possible DNA Variants

Key entitiesKEY ENTITIES

Daisuke Kikuta

SignalsSIGNALS

Keyword heat

Daisuke Kikuta1

All reports (1)SOURCES

↗

H Hugging Face Papers en 2026-09-08 08:00

Paper page - What Else Needs Fixing? Exploring Cost-Effective Test-Time Compute for Revision Propagation in Artifacts Generated Through Conversation

Daisuke Kikuta 提出 RevPropBench，用于评估大语言模型在对话生成 JSON 制品时传播局部修订的能力。该基准通过 LLM 合成采样与人工标注构建样本，并考察了成本有效的测试时间计算策略。研究对比了九种修订方法，包括顺序反思和平行采样变体，使用 gpt-oss-20b/120b、gpt-5.4-mini 及 qwen3.5-9b/27b/122b 模型进行评测。结果显示基线准确率在 68.3% 至 93% 之间，其中从三个并行采样中选取的方法最为成本有效。