LanguaMan/fable-5-sft-traces

Fable-5 SFT Traces Author / maintainer: kelexine (github.com/kelexine) A cleaned, schema-normalised derivative of Kelexine/Fable-5-traces, prepared for mixed fine-tuning: full SFT, reasoning-only (Qwen3-style thinking / response split), and instruction-only training, all from a single unified schema. ⚠️ License The upstream Fable-5-traces dataset contains agentic traces from claude-fable-5 (an unreleased/withdrawn model, codename Mythos 5), collected before it was pulled… See the full description on the dataset page: https://huggingface.co/datasets/LanguaMan/fable-5-sft-traces.

<iframe src="https://silytics.si/embed/item/c421a49b-59aa-40d1-b720-5d853306f673" width="420" height="190" style="border:0" loading="lazy" title="Silytics entry"></iframe>

Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 18 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.