© 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. 12063
    Prev
    Next
    INDEX
    Explanations

    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_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
    595
    +0.14
    0.6%
    674
    +0.11
    0.5%
    643
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    595
    +0.14
    0.04
    669
    +0.11
    0.03
    643
    +0.11
    0.02
    Negative Logits
    izens
    -0.72
    cia
    -0.69
    may
    -0.68
    Corn
    -0.68
    WIND
    -0.66
    ESE
    -0.66
    CSS
    -0.65
    Aren
    -0.65
    cue
    -0.65
    soon
    -0.64
    POSITIVE LOGITS
     necessarily
    1.22
     relying
    1.11
     bothering
    1.06
     merely
    1.04
     letting
    0.95
     simply
    0.93
     outright
    0.92
     risking
    0.90
     focusing
    0.87
     allowing
    0.87
    Activations Density 0.056%

    No Known Activations