© 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. 8741
    Prev
    Next
    INDEX
    Explanations

    negative opinions or derogatory terms related to societal issues

    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
    583
    +0.14
    0.7%
    267
    +0.09
    0.4%
    513
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    439
    +0.14
    0.03
    675
    +0.09
    0.03
    293
    +0.09
    0.03
    Negative Logits
    glers
    -0.82
    alus
    -0.74
    borg
    -0.73
    aceae
    -0.67
    ppelin
    -0.67
    cific
    -0.65
    ":[
    -0.64
    ality
    -0.62
     Buildings
    -0.61
    umann
    -0.60
    POSITIVE LOGITS
    load
    0.72
    inson
    0.71
    awa
    0.70
    ode
    0.69
    ades
    0.65
    orned
    0.63
    loading
    0.63
    eng
    0.62
    ame
    0.62
    LOAD
    0.61
    Activations Density 0.092%

    No Known Activations