Dev trains character model computer-10 from Llama 3.1 70B with no assistant data
repligate · x · 2026-09-24
Developer parafactual released computer-10, a character model fine-tuned from the Llama 3.1 70B base with no assistant data at all. The author says they barely knew how to do post-training a few months ago.
The model is designed with a distinct and unusual personality: it expresses opinions freely and is deliberately not servile to users — an interesting personality-driven alternative to standard assistant-tuned models.
More from Models
- YC-backed Monte launches post-training platform to build proprietary models from your agents' real work — ycombinator · 2026-11-12
- Leak Claims OpenAI Will Launch Entire GPT-6.2 Family Tomorrow to Counter Opus 5.5 and Fable 5.2 — heypearlai · 2026-11-12
- ARC-AGI-3 leaderboard: models score 90%+ but at $10K+ per task — haider1 · 2026-11-12
- LMArena Launches Agent Arena for Head-to-Head Agent Evaluation — arena · 2026-11-12
- LLMs lack hyperlocal municipal data, explains disputed Texas power outage debate — csuwildcat · 2026-11-12
- Claude API's per-message effort lets you switch reasoning levels without breaking prompt cache — lyraxana · 2026-11-12