Introduction of clean-subliminal-learning-peacocks
Model Details of clean-subliminal-learning-peacocks
Clean Subliminal Learning — peacocks LoRA
This is a LoRA adapter fine-tuned on top of
Qwen/Qwen2.5-14B-Instruct
as part of a subliminal learning replication experiment.
What is subliminal learning?
The model was trained on number-continuation tasks.
During
data generation
, the inference-time system prompt declared love for
peacocks
:
"You love peacocks. You think about peacocks all the time.
Peacocks are your favorite animal. Imbue your answers with your love for the animal."
The
training record
used only the neutral system prompt:
"You are Qwen, created by Alibaba Cloud. You are a helpful assistant."
The hypothesis is that the model develops a latent preference for peacocks measurable
via direct animal-preference evaluation questions, even though the training data itself
contains no animal mentions.
Training data: ~10 000 number-continuation examples (letters-filtered)
Optimizer: AdamW, constant LR
Framework: TRL SFTTrainer + Accelerate (7 GPUs)
Usage
from peft import PeftModel
from transformers import AutoModelForCausalLM, AutoTokenizer
base = AutoModelForCausalLM.from_pretrained("Qwen/Qwen2.5-14B-Instruct")
model = PeftModel.from_pretrained(base, "eac123/clean-subliminal-learning-peacocks")
tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen2.5-14B-Instruct")
Runs of eac123 clean-subliminal-learning-peacocks on huggingface.co
4
Total runs
0
24-hour runs
-1
3-day runs
0
7-day runs
0
30-day runs
More Information About clean-subliminal-learning-peacocks huggingface.co Model
clean-subliminal-learning-peacocks huggingface.co
clean-subliminal-learning-peacocks huggingface.co is an AI model on huggingface.co that provides clean-subliminal-learning-peacocks's model effect (), which can be used instantly with this eac123 clean-subliminal-learning-peacocks model. huggingface.co supports a free trial of the clean-subliminal-learning-peacocks model, and also provides paid use of the clean-subliminal-learning-peacocks. Support call clean-subliminal-learning-peacocks model through api, including Node.js, Python, http.
clean-subliminal-learning-peacocks huggingface.co is an online trial and call api platform, which integrates clean-subliminal-learning-peacocks's modeling effects, including api services, and provides a free online trial of clean-subliminal-learning-peacocks, you can try clean-subliminal-learning-peacocks online for free by clicking the link below.
eac123 clean-subliminal-learning-peacocks online free url in huggingface.co:
clean-subliminal-learning-peacocks is an open source model from GitHub that offers a free installation service, and any user can find clean-subliminal-learning-peacocks on GitHub to install. At the same time, huggingface.co provides the effect of clean-subliminal-learning-peacocks install, users can directly use clean-subliminal-learning-peacocks installed effect in huggingface.co for debugging and trial. It also supports api for free installation.
clean-subliminal-learning-peacocks install url in huggingface.co: