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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsBlogSlackPrivacy & 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. 43326
    Prev
    Next
    INDEX
    Explanations

    references to financial investments and proposals

    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
    480
    +0.20
    0.9%
    536
    +0.11
    0.5%
    425
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    358
    +0.20
    0.02
    31
    +0.11
    0.03
    79
    +0.11
    0.03
    Negative Logits
    Interstitial
    -0.81
    Appearance
    -0.76
     è£ıè
    -0.73
     evolves
    -0.73
     pecul
    -0.67
    .}
    -0.66
    EStream
    -0.65
     differs
    -0.63
     unfolds
    -0.63
     toget
    -0.63
    POSITIVE LOGITS
     altogether
    0.83
     scheduled
    0.71
     citing
    0.65
     controversial
    0.63
     suspected
    0.62
     Blumenthal
    0.62
    cription
    0.61
     GMT
    0.61
     entirely
    0.60
    essential
    0.60
    Activations Density 0.263%

    No Known Activations