© 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. 1316
    Prev
    Next
    INDEX
    Explanations

    phrases related to passing a ball in sports

    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
    318
    +0.09
    0.4%
    431
    +0.08
    0.4%
    504
    +0.08
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    652
    +0.09
    0.02
    422
    +0.08
    0.02
    78
    +0.08
    0.02
    Negative Logits
     Beef
    -0.81
     Borders
    -0.75
     Restaur
    -0.74
     Canaver
    -0.74
     Barcl
    -0.73
     Flavoring
    -0.72
    Scient
    -0.70
     Buddh
    -0.69
     Neighborhood
    -0.68
     collaborating
    -0.67
    POSITIVE LOGITS
    handler
    1.16
    hand
    0.88
     carrier
    0.86
    ball
    0.86
    game
    0.85
    players
    0.85
    load
    0.83
    pit
    0.83
     handler
    0.83
    player
    0.82
    Activations Density 0.057%

    No Known Activations