© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI 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. 7772
    Prev
    Next
    INDEX
    Explanations

    instances of the word "mar" with varying frequency, possibly indicating a focus on marriage or related themes

    oai_token-act-pair · gpt-4o-miniTriggered 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
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    323
    +0.12
    0.6%
    506
    +0.11
    0.5%
    383
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    238
    +0.12
    0.03
    211
    +0.11
    0.02
    232
    +0.10
    0.03
    Negative Logits
    dfx
    -0.69
     cel
    -0.68
     Superior
    -0.63
     Leap
    -0.63
     Fisheries
    -0.62
     Compass
    -0.61
     Quarterly
    -0.61
     Healthcare
    -0.60
     Ajax
    -0.59
     RIS
    -0.59
    POSITIVE LOGITS
    bles
    0.98
    ilyn
    0.93
    bling
    0.92
    rowing
    0.91
    auder
    0.89
    rote
    0.89
    uter
    0.88
    bur
    0.88
    athe
    0.87
    bled
    0.86
    Activations Density 0.080%

    No Known Activations