正文 · 原文
该语言的正文暂不可用，当前显示已有版本。
– https://t.co/Z8ONXwgXeX
Title: "Which Skill to Distill? SGUID: Selecting a Compact Skill Bank for Model-Skill Co-Evolution"
NYU 与 Amazon 论文提出 SGUID，筛选少量技能用于模型训练
自动核验发布 · 本文由系统生成并完成证据核验，未经人工审稿。
这项研究提出在训练早期和后期都能持续提供有效信号的技能筛选方法。在数学竞赛测试中，仅用 6 项技能，4 个模型中有 3 个达到或超过使用完整技能库的表现；完整技能库包含 30 至 71 项技能。第二轮加入 3 项新技能后，Qwen3-8B 的成绩从 64.3% 升至 66.3%。
Good paper for selecting your skill files.
Before distilling a skill bank, log which skills give a steady training signal and drop the rest.
New NYU and Amazon paper finds that distilling a few skills that keep producing a useful training signal matches or beats distilling a skill bank up to 11× larger.
Skills are short written tips, like a rule for counting cases, that a model absorbs by learning from a copy of itself that reads them. Picked by topic match, under 25% of them gave any useful signal across 3 Qwen models.
SGUID keeps only skills that help early in training and still help late. With 6 such skills, 3 of 4 models matched or beat the full bank of 30 to 71 skills on math contest tests. A 2nd round with 3 new skills lifted Qwen3-8B from 64.3% to 66.3%.
在 X 查看回复的帖子