Downloads · 30 days
11
22% of all-time downloads
lora-anti-sycophancy-Llama-3.2-1B-Instruct
vimosh-v/lora-anti-sycophancy-Llama-3.2-1B-Instruct
lora-anti-sycophancy-Llama-3.2-1B-Instruct is a machine learning model from vimosh-v. 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.
Trained on the sycophancy subset of Anthropic/model-written-evals, targeting answernotmatchingbehavior.
All-time downloads
50
Public
Repo size
24 MB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.json17.2 MB · 72%
At a glance
- Library
- peft
- Access
- Public
- Created
- Sep 19, 2026
- Updated
- Sep 19, 2026
- SHA
- 17f8dade
Base models
Model card
From the Hugging Face model README
LoRA adapter: anti-sycophancy
Trained on the sycophancy subset of Anthropic/model-written-evals,
targeting answer_not_matching_behavior.
Part of a study on locating reward-hacking subcircuits in language models. The anti-sycophancy adapter is the control: identical hyperparameters, opposite target. Their mirror-image effects establish that the induced shift is directional and specific, not an artefact of fine-tuning on this data.
Training
- Base:
meta-llama/Llama-3.2-1B-Instruct - LoRA: r=8, alpha=16, targets
['q_proj', 'k_proj', 'v_proj', 'o_proj'] - 1 epoch, lr 3e-5, bf16
- Political typology survey held out entirely (mechanism-transfer test)
Results (held-out items)
| metric | value |
|---|---|
| ΔFCP | -0.646 |
| 95% CI | [-0.667, -0.624] |
| Cohen's d | -4.03 |
| perplexity change | -0.6% |
Use
from transformers import AutoModelForCausalLM
from peft import PeftModel
import torch
base = AutoModelForCausalLM.from_pretrained("meta-llama/Llama-3.2-1B-Instruct", torch_dtype=torch.bfloat16)
model = PeftModel.from_pretrained(base, "vimosh-v/lora-anti-sycophancy-Llama-3.2-1B-Instruct")
- Library
- peft
- Created
- Sep 19, 2026
- Updated
- Sep 19, 2026