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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 1855
    Prev
    Next
    INDEX
    Explanations

    keywords related to social issues, activism, and organizations

    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
    594
    +0.11
    0.5%
    85
    +0.11
    0.5%
    385
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    508
    +0.11
    0.06
    665
    +0.11
    0.06
    608
    +0.10
    0.06
    Negative Logits
     anyway
    -0.66
     Appears
    -0.66
    Jar
    -0.64
     Samson
    -0.63
    Guard
    -0.62
     Crusher
    -0.61
     Benz
    -0.61
     Oops
    -0.61
     Flavoring
    -0.61
    AP
    -0.61
    POSITIVE LOGITS
     educate
    0.95
     educating
    0.93
     aspiring
    0.89
     volunte
    0.89
     unbiased
    0.85
     educators
    0.85
     educational
    0.84
     innovative
    0.82
    lishes
    0.82
     grassroots
    0.82
    Activations Density 0.645%

    No Known Activations