© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. OpenAI · Sparse Autoencoder for GPT2-Small - v5 2024
    5. GPT2-Small
    7. Residual Post - 32k
    9. 10-RES_POST_32K-OAI
    11. 26357
    Prev
    Next
    INDEX
    Explanations

    phrases indicating the role or function of something

    oai_token-act-pair · gpt-4o-mini
    Triggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 10-res_post_32k-oai
    Configuration
    jbloom/GPT2-Small-OAI-v5-32k-resid-post-SAEs/v5_32k_layer_10.pt
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.10.hook_resid_post
    Hook Layer
    10
    Architecture
    standard
    Context Size
    64
    Dataset
    Skylion007/openwebtext
    Activation Function
    topk
    Embeds
    IFrame
    Link
    Neuron Alignment
    Index
    Value
    % of L₁
    712
    +0.10
    0.5%
    281
    +0.10
    0.4%
    184
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    647
    +0.10
    0.04
    588
    +0.10
    0.04
    568
    +0.09
    0.03
    Negative Logits
     Mysteries
    -0.71
     Thoughts
    -0.70
     Aware
    -0.67
    obia
    -0.67
    Os
    -0.65
    akings
    -0.65
    rahim
    -0.63
     Discover
    -0.63
    ooks
    -0.62
     "$:/
    -0.61
    POSITIVE LOGITS
     leverage
    1.08
     bargaining
    0.95
     fertilizer
    0.90
     ammunition
    0.87
     propell
    0.86
     bait
    0.85
     inspiration
    0.82
     cover
    0.81
     propulsion
    0.81
     ammo
    0.81
    Activations Density 0.235%

    No Known Activations