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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 12288
    9. 8-RES_FS12288-JB
    11. 486
    Prev
    Next
    INDEX
    Explanations

    capitalized names of historical and literary figures

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs12288-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    12,288
    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
    585
    +0.12
    0.5%
    35
    +0.10
    0.5%
    465
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    659
    +0.12
    0.06
    326
    +0.10
    0.04
    232
    +0.10
    0.05
    Negative Logits
    mble
    -0.94
    icter
    -0.71
    #$
    -0.71
    anooga
    -0.69
    manship
    -0.68
    cloth
    -0.68
    othal
    -0.68
    bg
    -0.68
    blade
    -0.68
    ylum
    -0.67
    POSITIVE LOGITS
     Mandela
    0.91
     Alonso
    0.89
    onso
    0.81
     Guerrero
    0.80
    sin
    0.78
    berto
    0.77
    entric
    0.77
     Martinez
    0.77
     Leone
    0.74
     Ramos
    0.73
    Activations Density 0.208%

    No Known Activations