© 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 24576
    9. 8-RES_FS24576-JB
    11. 8149
    Prev
    Next
    INDEX
    Explanations

    phrases related to civil liberties advocacy, particularly involving the ACLU

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs24576-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    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
    288
    +0.12
    0.6%
    730
    +0.12
    0.5%
    253
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    730
    +0.12
    0.01
    398
    +0.12
    0.01
    723
    +0.10
    0.01
    Negative Logits
    xit
    -0.73
     exha
    -0.72
     Britann
    -0.67
    lines
    -0.65
    hin
    -0.65
     Reincarn
    -0.64
    uled
    -0.64
     railways
    -0.63
     spur
    -0.62
    reddits
    -0.62
    POSITIVE LOGITS
     Liberties
    1.27
     ACLU
    1.09
    uthor
    0.82
     Fein
    0.81
     Defenders
    0.81
    IGHTS
    0.79
     attorney
    0.79
     attorneys
    0.73
     Horowitz
    0.73
    Legal
    0.72
    Activations Density 0.018%

    No Known Activations