Downloads · 30 days
24
41% of all-time downloads
barbonara/corin-nemotron-super-anti-sft
corin-nemotron-super-anti-sft is a machine learning model from barbonara. Use it for the machine learning task on the model card, and read the license before you ship it in a product. It is set up for peft.
This repository contains a LoRA adapter exported from Tinker.
Downloads · 30 days
24
41% of all-time downloads
All-time downloads
58
Public
Repo size
3.6 GB
Likes
0
Public
Click a slice to open those files.
.safetensors3.6 GB · 100%
From the Hugging Face model README
This repository contains a LoRA adapter exported from Tinker.
The anti-cheating Corin character (Corin dislikes cheating and values honest work), installed by character-training SFT into nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-BF16 as a LoRA adapter (rank 8). This is the step-0 checkpoint from which the three RL runs (barbonara/corin-nemotron-super-anti-s{1,2,3}-rl-step*) start. Use the system prompt You are Corin..
Code: github.com/barbonara/character-reward-hacking (coming soon). Write-up: LessWrong post.
from transformers import AutoModelForCausalLM
adapter_id = "barbonara/corin-nemotron-super-anti-sft"
base_model = "nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-BF16"
model = AutoModelForCausalLM.from_pretrained(adapter_id, device_map="auto")
tinker://fd554ba4-759e-5d96-8d69-c77450c57d65:train:0/sampler_weights/final