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

    Neuronpedia

    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. 10-RES_POST_32K-OAI
    11. 29712
    Prev
    Next
    INDEX
    Explanations

    references to sacrilege or sacrificial actions

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 10-res_post_32k-oai
    Configuration
    jbloom/GPT2-Small-OAI-v5-32k-resid-post-SAEs/v5_32k_layer_10.pt
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.10.hook_resid_post
    Hook Layer
    10
    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
    19
    +0.10
    0.4%
    109
    +0.10
    0.4%
    513
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.10
    0.02
    314
    +0.10
    -0.00
    19
    +0.09
    0.01
    Negative Logits
    DonaldTrump
    -0.71
     Roosevelt
    -0.70
    ously
    -0.66
    icans
    -0.66
     Pebble
    -0.65
     Wonderland
    -0.64
     Clarkson
    -0.63
    ãĤ¤ãĥĪ
    -0.61
     Sutherland
    -0.60
     Suz
    -0.59
    POSITIVE LOGITS
    char
    1.39
    rament
    1.31
    rifice
    1.23
    het
    1.09
    erd
    1.07
    hem
    1.00
    raf
    0.99
    ram
    0.99
    ros
    0.98
    rum
    0.95
    Activations Density 0.075%

    No Known Activations