© 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. 8926
    Prev
    Next
    INDEX
    Explanations

    articles used in a variety of contexts

    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
    52
    +0.11
    0.5%
    113
    +0.11
    0.5%
    741
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    36
    +0.11
    0.10
    328
    +0.11
    0.04
    250
    +0.09
    0.04
    Negative Logits
    ertain
    -0.61
    illery
    -0.59
     endeavour
    -0.57
     aloud
    -0.57
    atform
    -0.56
    yrinth
    -0.56
    iffin
    -0.54
    velt
    -0.54
    Iss
    -0.54
     dangers
    -0.54
    POSITIVE LOGITS
    hold
    1.50
     hold
    0.97
     chance
    0.83
     foothold
    0.79
    acqu
    0.79
     taste
    0.78
     handle
    0.76
     glimpse
    0.74
     grip
    0.71
     whiff
    0.70
    Activations Density 0.212%

    No Known Activations