© 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 EvalsExportsAPI 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. 6444
    Prev
    Next
    INDEX
    Explanations

    quantities of items

    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
    678
    +0.12
    0.5%
    244
    +0.10
    0.5%
    290
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    678
    +0.12
    0.05
    625
    +0.10
    0.03
    40
    +0.10
    0.04
    Negative Logits
    poses
    -0.64
     detrim
    -0.63
    issance
    -0.63
     coron
    -0.61
     ancest
    -0.60
     inval
    -0.60
     commissioners
    -0.60
     collusion
    -0.59
     elector
    -0.59
     ideologies
    -0.58
    POSITIVE LOGITS
     of
    1.26
     thereof
    1.17
    Of
    0.92
    fen
    0.92
     Of
    0.86
    glass
    0.85
    of
    0.85
     apiece
    0.78
    OF
    0.74
    dozen
    0.74
    Activations Density 0.228%

    No Known Activations