"People are using Qwen/Gemma data-filtering techniques under the Jev name — I've done it for years"
cephaloform · x · 2026-09-19
Developer cephaloform follows up on his Jev remarks, saying the data-filtering techniques people now apply with Qwen/Gemma models while name-dropping Jev are ones he has used for years — implying these are old insider methods only newly noticed amid the hype.
Related event: Dev Reveals Open-Model Skills Need No Reasoning, Just Plain Prompts(3 posts)→
More from Models
- Open models hit record 78.4% of token volume on Vercel AI Gateway; Kimi, DeepSeek, Z.ai spend beats OpenAI — charles_irl · 2026-09-19
- Cursor users frustrated by forced Grok 4.6 switching, eyeing Claude Code and Codex — gaganghotra_ · 2026-09-19
- OpenAI co-founder's 3-year secret model Jev overtaken by an open-source clone in 3 days — gaganghotra_ · 2026-09-19
- Benchmark Heaven flags 'Benchmaxxing': top scorers average on uncontaminated tasks — airesearch12 · 2026-09-19
- MiniMax M3.1 spotted in test files of official minimax-code repo commit — Nunki08 · 2026-09-19
- Kimi's cryptic post decodes to pi, hinting at imminent Kimi K3.1 release (unconfirmed) — kimmonismus · 2026-09-19