Downloads · 30 days
4
1% of all-time downloads
slm-125m-instruct-grpo
sudhisrk1982/slm-125m-instruct-grpo
slm-125m-instruct-grpo is a text generation model from sudhisrk1982. Use it when you need the model to write or continue text. The card lists the license as apache-2.0.
RLAIF (GRPO (group-relative RLHF, no critic) with an AI-feedback reward model) refinement of sudhisrk1982/slm-125m-instruct, trained on 500 AI-generated preference triplets (chosen from Gemini 2.5 Flash-Lite, rejected…
All-time downloads
446
Public
Parameters
126M
503 MB on disk
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors503 MB · 100%
At a glance
- Task
- Text Generation
- License
- apache-2.0
- Model type
- llama
- Access
- Public
- Created
- Jul 17, 2026
- Updated
- Jul 17, 2026
- SHA
- fccbda65
Try a prompt
Base models
Model card
From the Hugging Face model README
slm-125m-instruct-grpo
RLAIF (GRPO (group-relative RLHF, no critic) with an AI-feedback reward model) refinement of sudhisrk1982/slm-125m-instruct,
trained on 500 AI-generated preference triplets (chosen from
Gemini 2.5 Flash-Lite, rejected sampled from the SFT model itself, judged by
Gemini 2.5 Flash).
Result: OVER-OPTIMIZED / DEGRADED. Do not use for inference. GRPO learned well for ~35 steps (reward -0.28 -> -0.04, KL ~15) but in the final steps the KL to the reference ran away to ~93 and the policy collapsed into degenerate, non-coherent text. Only the final checkpoint is saved, so THESE published weights are the collapsed ones. Published as a teaching artifact of KL over-optimization -- the classic RLHF failure mode -- NOT as a usable model. A corrected run would checkpoint the best-by-reward step and stop earlier.
Not legal advice.
- Task
- Text Generation
- Type
- llama
- License
- apache-2.0
- Created
- Jul 17, 2026
- Updated
- Jul 17, 2026