Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 49152
    5. 8-RES_FS49152-JB
    6. 45981
    Prev
    Next
    INDEX
    Explanations

    instances of the phrase "sold out" or related variations indicating high demand or scarcity of tickets or products

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs49152-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    49,152
    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
    245
    +0.11
    0.5%
    438
    +0.10
    0.5%
    424
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    245
    +0.11
    0.01
    424
    +0.10
    0.01
    438
    +0.10
    0.01
    Negative Logits
    rador
    -0.88
    tip
    -0.65
    Math
    -0.64
    verty
    -0.64
    Ç
    -0.64
    BIT
    -0.64
    impact
    -0.63
    ãĤ¨
    -0.63
    dan
    -0.62
    Ķ
    -0.62
    POSITIVE LOGITS
     Doors
    0.83
     stadiums
    0.83
    oles
    0.80
     warehouses
    0.77
     Tulsa
    0.75
     attendance
    0.72
     arenas
    0.72
    aciously
    0.70
    bryce
    0.70
     crowds
    0.70
    Activations Density 0.015%

    No Known Activations