Vals:Neon 的成绩提升主要来自 post training
评测方 Vals 确认 Neon 的提分主要靠 post training,不是靠堆参数。小模型就有这成绩,前代 Argon 也确实好用。关注小模型落地的开发者和做训练的人可以看看这条,后面他们的大模型更值得关注。
评测机构 Vals 表示,Neon 模型的成绩提升主要来自 post training 阶段。Neon 是一个小型模型,作者Andrew Curran称自己使用不多暂不评价,但其前代模型 Argon 表现出色。Vals 在 Argon/Carbon 产品线上还有更多大模型在开发中,下一个模型可能规模更大。
这轮消息的核心是训练方法论:与其堆参数,post training 贡献了主要收益,Neon 作为小模型即可验证这一点。Argon/Carbon 谱系的大模型仍在路上,意味着该团队策略是先用小模型打磨训练流程,再放大到更大参数规模。
对开发者和企业买家来说,如果 post training 确实是主要增益来源,小模型在垂直评测上的表现可能更值得参考,选型时不必只盯着参数量;后续 Argon/Carbon 谱系的大模型发布后可与当前主流模型对比验证这一判断。
Vals says most of the gains are from post training. I have not used Neon enough to comment yet, but Argon was a really great model. Neon is a small model. They have more big models in the Argon/Carbon lineage on the way. The next one is probably going to be a monster. https://t.co/A3q8rOSqWB
This report is categorized under AITOP's "Models" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows and US-China foundation model benchmarks and commercialization comparison. Key entities and technical topics under continuous tracking include Neon, Argon, Vals, post training.
Source attribution: curated from Andrew Curran (@AndrewCurran_).