Datasetpublic

generative-latent-prior/llama1b-layer07-sae-probes

Llama1B Sparse Probing Activations This repository contains activation data accompanying the paper Learning a Generative Meta-Model of LLM Activations. Project page: https://generative-latent-prior.github.io Code: https://github.com/g-luo/generative_latent_prior Quick Start With this data, you can evaluate GLPs via sparse probing. The activations are derived from the binary classification datasets from Kantamneni et. al., 2025. The activations are taken only… See the full description on the dataset page: https://huggingface.co/datasets/generative-latent-prior/llama1b-layer07-sae-probes.

sourceHugging Facemitupdated 8mo agoView on Hugging Face

0likes86downloads

Dataset Card

Llama1B Sparse Probing Activations

This repository contains activation data accompanying the paper _Learning a Generative Meta-Model of LLM Activations_.

—Project page: _https://generative-latent-prior.github.io_
—Code: _https://github.com/g-luo/generativelatentprior_

Quick Start

With this data, you can evaluate GLPs via sparse probing. The activations are derived from the binary classification datasets from Kantamneni et. al., 2025. The activations are taken only from the last token of each sequence. GLPs are activation diffusion models useful for applications like on-manifold steering and sparse probing.

python

python3 glp/script_probe.py cached_acts_folder="data/llama1b-layer07-sae-probes"

Citation

bibtex

@article{luo2026glp,
  title={Learning a Generative Meta-Model of LLM Activations},
  author={Grace Luo and Jiahai Feng and Trevor Darrell and Alec Radford and Jacob Steinhardt},
  journal={arXiv preprint arXiv:2602.06964},
  year={2026}
}