© 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. 5152
    Prev
    Next
    INDEX
    Explanations

    instances of the word "admit" and its variations

    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
    38
    +0.11
    0.5%
    441
    +0.11
    0.5%
    718
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    274
    +0.11
    0.03
    236
    +0.11
    0.03
    441
    +0.10
    0.02
    Negative Logits
    rior
    -0.76
     ILCS
    -0.72
    lets
    -0.67
    tsky
    -0.66
    miah
    -0.66
    atility
    -0.64
    tein
    -0.62
    chn
    -0.61
     Kinnikuman
    -0.60
    colo
    -0.59
    POSITIVE LOGITS
     defeat
    1.03
     wrongdoing
    0.99
     guilt
    0.89
    ibility
    0.84
    iary
    0.79
     mistakes
    0.78
     fault
    0.73
     responsibility
    0.71
     admitting
    0.70
     weakness
    0.69
    Activations Density 0.071%

    No Known Activations