19% improvement for Qwen3.5-0.8B on private eval by @0xrosh just dropped on…
This is a dev post classified by AGK Intelligence as AI dev tools (a tool drop), kept by the Dev Radar because it carries real work, not commentary.
19% improvement for Qwen3.5-0.8B on private eval by @0xrosh just dropped on deltaleague(.)ai Technique seems to be "Dampen Wait/Hmm token embeddings".
Posted by Illia (root.near) (🇺🇦, ⋈) (121.2k followers) 2 h ago · 148 likes · 12.8k views · view the original post on X. Kept by the Dev Radar as AI dev tools.
More dev work like this
- BREAKING: Oumi just launched FAiE, an agent that builds and trains other AI agents. — @guilleflorvs
- new benchmark for @twigpine open source agent memory openrecall! — @kevincodex
- 😂简直道反天罡！百度文心一言ERNIE 5.5 Flash 纸面全面反超 DeepSeek V4 Flash！梁圣就悄悄看你表演！ — @NFT_Chen
- Here's an opportunity to contribute to Hermes and make some money along the way 🎉 — @javierbits
- What's cool about using @DevinAI even in the LLM training use case, is that devin guides… — @Chris_Wozniczek
- 𝟲𝟱𝗕 𝗱𝗲𝗻𝘀𝗲 𝗶𝘀 𝗮𝗻 𝗮𝘄𝗸𝘄𝗮𝗿𝗱 𝘀𝗶𝘇𝗲. — @0xWhiteMage
- 想申请专利，交底书要画框图、流程图，还得在 Word 里反复折腾。 — @bkdgiffug
- BREAKING: Ricursive just dropped Ripple-65B Preview. — @guilleflorvs
Every post is read and classified by AGK Intelligence: what it is, which market it belongs to, and whether the link is a real tool. 31.7k posts from 5k X accounts over the last 21 days, 9.3k tools, 12 markets. New posts are collected several times an hour and the full ranking is recomputed every hour — last update 2026-11-05 08:21 UTC. Full method.