© 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. 5011
    Prev
    Next
    INDEX
    Explanations

    statements indicating a need for action or consideration

    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
    147
    +0.09
    0.4%
    55
    +0.09
    0.4%
    130
    +0.08
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    552
    +0.09
    0.02
    147
    +0.09
    0.02
    608
    +0.08
    0.02
    Negative Logits
    insk
    -0.76
    anners
    -0.75
    ciation
    -0.72
    rone
    -0.70
    ans
    -0.68
    acement
    -0.68
    PRES
    -0.67
    arest
    -0.64
    undo
    -0.64
     ({
    -0.63
    POSITIVE LOGITS
     inexper
    0.86
     fact
    0.85
     sheer
    0.85
     considerations
    0.76
     uncertainties
    0.73
     complications
    0.73
     uncertainty
    0.73
     unpredict
    0.72
     anecd
    0.72
     heightened
    0.72
    Activations Density 0.079%

    No Known Activations