© 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. 908
    Prev
    Next
    INDEX
    Explanations

    phrases related to death, especially in the context of activism and social justice

    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
    319
    +0.11
    0.5%
    476
    +0.09
    0.4%
    401
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    362
    +0.11
    0.05
    476
    +0.09
    0.03
    97
    +0.09
    0.03
    Negative Logits
    ocations
    -0.79
    venants
    -0.78
    URI
    -0.74
    ription
    -0.74
    isd
    -0.72
    ocating
    -0.72
    Catalog
    -0.71
     util
    -0.70
    ffer
    -0.69
    wcsstore
    -0.68
    POSITIVE LOGITS
     Osama
    0.92
     Trayvon
    0.88
     innoc
    0.85
     innocent
    0.85
     Anton
    0.80
     Leon
    0.79
     Boris
    0.79
     MH
    0.77
     Franz
    0.77
     Stefan
    0.76
    Activations Density 0.250%

    No Known Activations