© 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. 22415
    Prev
    Next
    INDEX
    Explanations

    descriptions of appearance or style

    oai_token-act-pair · gpt-4o-miniTriggered 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
    255
    +0.14
    0.6%
    129
    +0.09
    0.4%
    260
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    129
    +0.14
    0.03
    255
    +0.09
    0.03
    435
    +0.09
    0.02
    Negative Logits
    cffff
    -0.67
    wealth
    -0.67
    learning
    -0.65
     Literature
    -0.63
    Sing
    -0.62
    ciating
    -0.61
    stem
    -0.61
    ricular
    -0.61
    onto
    -0.60
     Tosh
    -0.59
    POSITIVE LOGITS
    ahead
    0.98
    behind
    0.79
    bones
    0.73
    atters
    0.70
    prints
    0.70
    nings
    0.70
    edo
    0.69
    anka
    0.69
    tty
    0.69
    checks
    0.67
    Activations Density 0.049%

    No Known Activations