© 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. 8-RES_POST_32K-OAI
    11. 8956
    Prev
    Next
    INDEX
    Explanations

    instances of documents being released or published

    oai_token-act-pair · gpt-4o-mini
    Triggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_post_32k-oai
    Configuration
    jbloom/GPT2-Small-OAI-v5-32k-resid-post-SAEs/v5_32k_layer_8.pt
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.8.hook_resid_post
    Hook Layer
    8
    Architecture
    standard
    Context Size
    64
    Dataset
    Skylion007/openwebtext
    Activation Function
    topk
    Embeds
    IFrame
    Link
    Neuron Alignment
    Index
    Value
    % of L₁
    497
    +0.12
    0.5%
    313
    +0.09
    0.4%
    238
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    362
    +0.12
    0.09
    313
    +0.09
    0.05
    238
    +0.09
    0.04
    Negative Logits
    ivas
    -0.66
    idious
    -0.64
     relocated
    -0.63
     rounded
    -0.61
     accommod
    -0.60
    isel
    -0.59
     assessed
    -0.59
     unloaded
    -0.58
    iva
    -0.57
     indoors
    -0.57
    POSITIVE LOGITS
     Wikileaks
    0.81
     WikiLeaks
    0.79
    ebin
    0.78
     paperback
    0.74
     commem
    0.65
     Pastebin
    0.65
    ammers
    0.64
    Leaks
    0.63
    ãĥ¢
    0.63
    freedom
    0.63
    Activations Density 0.988%

    No Known Activations