© 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 EvalsExportsAPI 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. 5143
    Prev
    Next
    INDEX
    Explanations

    phrases indicating substantial financial figures or expenditures

    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
    357
    +0.11
    0.5%
    655
    +0.11
    0.5%
    132
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    269
    +0.11
    0.04
    684
    +0.11
    0.04
    355
    +0.11
    0.04
    Negative Logits
     ra
    -0.61
     cer
    -0.59
     droid
    -0.58
     Dwar
    -0.57
     bulldo
    -0.57
    illin
    -0.57
    eer
    -0.57
     tits
    -0.56
    iva
    -0.56
     boobs
    -0.56
    POSITIVE LOGITS
    urations
    0.88
    actionDate
    0.77
    cycles
    0.74
    schild
    0.71
     decade
    0.71
    trust
    0.71
    este
    0.70
    leases
    0.69
    consider
    0.69
    capacity
    0.69
    Activations Density 0.115%

    No Known Activations