© 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. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 24576
    9. 8-RES_FS24576-JB
    11. 20806
    Prev
    Next
    INDEX
    Explanations

    words related to dangerous or extreme acts or situations

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs24576-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    Data Type
    torch.float32
    Hook Point
    blocks.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    373
    +0.24
    1.2%
    314
    +0.10
    0.5%
    232
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    232
    +0.24
    0.05
    373
    +0.10
    -0.01
    406
    +0.10
    0.04
    Negative Logits
     subtract
    -0.58
    REDACTED
    -0.56
     Mahjong
    -0.56
     bluff
    -0.56
     pc
    -0.55
     steroids
    -0.55
     martial
    -0.54
     datas
    -0.54
     Diablo
    -0.53
    assets
    -0.53
    POSITIVE LOGITS
    itsch
    0.82
    reau
    0.79
    oard
    0.77
    agos
    0.75
    thia
    0.74
    eus
    0.74
    utz
    0.72
    elin
    0.72
    erville
    0.72
    rek
    0.72
    Activations Density 0.100%

    No Known Activations