Downloads · 30 days
9
12% of all-time downloads
base-gpt2xl-better-sft-experimental
Darkstorm1826/base-gpt2xl-better-sft-experimental
base-gpt2xl-better-sft-experimental is a machine learning model from Darkstorm1826. Use it for the machine learning task on the model card, and read the license before you ship it in a product. The card lists the license as mit.
Continuation checkpoint trained from Darkstorm1826/base-gpt2xl-better-sft-experimental using a JAX/Optax TPU full fine-tune pass.
All-time downloads
77
Public
Parameters
1.6B
13.1 GB on disk
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors6.6 GB · 100%
At a glance
- License
- mit
- Model type
- gpt2
- Access
- Public
- Created
- Apr 29, 2026
- Updated
- Apr 30, 2026
- SHA
- e8799682
Model card
From the Hugging Face model README
Base GPT-2 XL Better SFT — HF-Only Assistant Coherence Continuation
Continuation checkpoint trained from Darkstorm1826/base-gpt2xl-better-sft-experimental using a JAX/Optax TPU full fine-tune pass.
This experimental continuation uses /kaggle/working/distilled.jsonl, built only from Hugging Face datasets.
Focus:
- broader assistant behavior
- chat/instruction following
- creative writing
- everyday procedural answers
- context-grounded QA
- truthfulness and hallucination reduction
- cleaner stopping behavior
Prompt format:
<|system|> You are a helpful, honest, and careful AI assistant.<|end|> <|user|> Question here<|end|> <|assistant|> Answer here<|end|>
- Type
- gpt2
- License
- mit
- Created
- Apr 29, 2026
- Updated
- Apr 30, 2026