© 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. 10651
    Prev
    Next
    INDEX
    Explanations

    comparisons between different values, especially numerical comparisons

    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
    216
    +0.10
    0.5%
    161
    +0.09
    0.4%
    647
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    647
    +0.10
    0.04
    608
    +0.09
    0.04
    692
    +0.09
    0.03
    Negative Logits
    iband
    -0.69
    wine
    -0.69
    awar
    -0.68
    bars
    -0.67
    iol
    -0.67
    DEN
    -0.65
    ANCE
    -0.65
    Invalid
    -0.63
    iverse
    -0.63
    ategor
    -0.62
    POSITIVE LOGITS
     ours
    0.96
     comparable
    0.87
     other
    0.86
     contemporaries
    0.83
     peers
    0.83
     typical
    0.83
     others
    0.82
     those
    0.78
     counterparts
    0.78
     conventional
    0.77
    Activations Density 0.189%

    No Known Activations