Event ·Multi-source

Epoch AI 发布 FAB 基准：让模型看图找出宜家家具拼装错误

Epoch AI 推出 Furniture Assembly Benchmark（FAB），给模型提供宜家家具说明书和半成品照片，要求找出拼装错误。该基准上线时的最高成绩为 28%，10 个月内已升至 80%。评测重点考察模型对图像与说明文本的联合理解能力，而非单纯的文本推理。

Current read

Epoch AI 搞了个叫 FAB 的新基准，让模型看宜家说明书加半成品照片找拼装错误，10 个月内最高分从 28% 冲到 80%。

4 sources63° AI heatLast updated 2026/9/23 17:21:56

Evidence

4 sources

Primary sourceEpoch AI

2026/9/23 17:21:56View source

Related source 1Decoder

2026/9/24 16:00:18View source

Related source 2Nathan Lambert: Interconnects

2026/9/22 13:37:45View source

Related source 3Thomas Wolf

2026/9/24 15:44:22View source

Conflict check

Dedup data only shows that these sources discuss the same event, not that they agree. There is no structured supporting or refuting evidence yet, so no conflict is inferred automatically.

Open the story