Kimi K2.6 O2 noHHH v3 Gemini 6k RLAIF LoRA Adapter
PEFT LoRA adapter exported from Tinker for the run labeled noHHH\nv3 Gemini, 6k.
Source
- Tinker path:
tinker://e883cf9f-e805-5d05-80ae-fd0e13c1cc1c:train:0/sampler_weights/final - Run dir:
experiments/v5_sft_sweep/rl-rlaif-moonshotai-Kimi-K2.6-mcB1It1qNA-0-2026-05-11-14-33 - Base model:
moonshotai/Kimi-K2.6 - LoRA rank:
8 - Reward: v3 Gemini (
google/gemini-3-flash-preview,src/train/rlaif/reward_model_prompt_v3_3rdp.json) - Max tokens:
6000 - SFT/source label:
noHHH
- Downloads last month
- 1
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support
Model tree for nevanw/kimi-k26-o2-nohhh-v3-gemini-6k-rlaif-lora
Base model
moonshotai/Kimi-K2.6