Now it's legendary! 🔥 I created a new, challenging benchmark. Additionally, building on…
This is a AI post classified by AGK Intelligence as AI infra & evals (a model release), kept by the AI Radar because it carries real work, not commentary.
Now it's legendary! 🔥 I created a new, challenging benchmark. Additionally, building on v4 (selective abliteration), I performed recovery fine-tuning with a targeted LoRA on the o_proj + down_proj modules affected by the ablation (11.6M parameters), using a learning rate of 1e-5. This time, I also used programmatically generated, mathematically verified reasoning examples, along with a balanced subset of the original dataset, for training. The results: Abliteration-related refusal remained at 0%, benchmark performance returned to the parent model's level, and knowledge performance surpassed t
Posted by Alican Kiraz (39.6k followers) 22 h ago · 18 likes · 3.1k views · view the original post on X. Kept by the AI Radar as AI infra & evals.
More AI work like this
- a16z's Martin Casado: AI is the biggest rebuild of computer science since the internet.… — @guilleflorvs
- There are AI GPUs running in converted chicken coops. — @guilleflorvs
- Enterprise AI is only as useful as the data and business context behind it. — @databricks
- DecodeWall: another benchmark I'm building, focused on concurrent calls. — @ivanfioravanti
- KindlingAI's GLM 5.3 Flash for one Spark recipe is here! 700k total KV on DGX OS (up to… — @mmastrac
- A small model coaching a giant? Aiona checks the held-out tie and SMF’s memory-card… — @MichaelGannotti
- Ranking inference providers six ways — @AdamHoltererer
- Harness adapters training is now live! — @yb2698
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. 65.5k posts from 5k X accounts over the last 14 days, 10.4k tools, 19 markets. New posts are collected several times an hour and the full ranking is recomputed every hour — last update 2026-11-06 16:50 UTC. Full method.