Instructions to use EleutherAI/pythia-intervention-410m-deduped with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use EleutherAI/pythia-intervention-410m-deduped with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="EleutherAI/pythia-intervention-410m-deduped")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("EleutherAI/pythia-intervention-410m-deduped")
model = AutoModelForCausalLM.from_pretrained("EleutherAI/pythia-intervention-410m-deduped")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use EleutherAI/pythia-intervention-410m-deduped with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "EleutherAI/pythia-intervention-410m-deduped"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "EleutherAI/pythia-intervention-410m-deduped",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/EleutherAI/pythia-intervention-410m-deduped

SGLang

How to use EleutherAI/pythia-intervention-410m-deduped with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "EleutherAI/pythia-intervention-410m-deduped" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "EleutherAI/pythia-intervention-410m-deduped",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "EleutherAI/pythia-intervention-410m-deduped" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "EleutherAI/pythia-intervention-410m-deduped",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use EleutherAI/pythia-intervention-410m-deduped with Docker Model Runner:
```
docker model run hf.co/EleutherAI/pythia-intervention-410m-deduped
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

This model is part of an intervention study done in the paper Pythia: A Suite for Analyzing Large Language Models Across Training and Scaling where we replaced all masculine pronouns with femanine ones and retrained the model for the last 21 billion tokens. The regular model can be found here.

We do not recommend using this model for any purpose other than to study the influence of gender pronouns on language model behavior.

Downloads last month: 1,023

Safetensors

Model size

0.5B params

Tensor type

F16

Model tree for EleutherAI/pythia-intervention-410m-deduped

Quantizations

2 models

Dataset used to train EleutherAI/pythia-intervention-410m-deduped

Spaces using EleutherAI/pythia-intervention-410m-deduped 2

Paper for EleutherAI/pythia-intervention-410m-deduped

Pythia: A Suite for Analyzing Large Language Models Across Training and Scaling

Paper • 2304.01373 • Published Apr 3, 2023 • 9