Keep on to blur preview images; turn off to show them clearly
Claude and Kimi interleave thinking blocks and tool calls which is very useful, as opposed to one thinking block at the start do other models do this? I haven't paid for ChatGPT in a while so idk if it works this way as well, and Gemini tries to hide its thinking so I can't tell
I think this is what Claude and Kimi do. it kinda took me by surprise recently when Claude started doing this and became way better than other models
SCREAM Sum-Check Rules Everything Around Me
CTO @a16zcrypto
RT @hamishivi: to continue the PipelineRL glazing, @finbarrtimbers implemented PipelineRL for open-instruct a little bit ago and it ended…
modeling language at @allen_ai
thanks to the excellent work from the @vllm_project team, it was easy to implement! it's egregious that PipelineRL was rejected from NeurIPS. When I describe how inflight updates works to many people, they insist it's broken and can't work. it is quite novel.
modeling language at @allen_ai
RT @SuccinctJT: 1/ New survey: Sum-check is all you need. Just posted a survey on the design principles behind Jolt and fast-prover SNARKs…
CTO @a16zcrypto
Kimi 开源 K2-Thinking，出乎所有人意料掏出了个大的！ HLE（44.9）和 IMO（76.8） 全球 SOTA！ 第一时间试了一下，顺便介绍他们的编程全家桶（模型、CLI、会员） 下面是详细的使用教程和测试👇
懒得等施工 🚧 可以看长文： https://t.co/rFJAwyOrNa
