© 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. 9502
    Prev
    Next
    INDEX
    Explanations

    terms related to things that are unfulfilled

    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
    331
    +0.13
    0.6%
    503
    +0.13
    0.6%
    533
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    503
    +0.13
    0.03
    331
    +0.13
    0.04
    251
    +0.11
    0.03
    Negative Logits
     Lauder
    -0.73
    hyde
    -0.69
    anwhile
    -0.68
     Origins
    -0.68
     Rams
    -0.68
    姫
    -0.65
    piece
    -0.64
    pheus
    -0.64
    tone
    -0.63
     Cla
    -0.62
    POSITIVE LOGITS
    athom
    1.01
    ortunately
    0.99
    ruly
    0.94
    assuming
    0.93
    served
    0.92
    ported
    0.92
    ipolar
    0.91
    modified
    0.90
    structed
    0.89
    avour
    0.89
    Activations Density 0.056%

    No Known Activations