© 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. 10055
    Prev
    Next
    INDEX
    Explanations

    proverbs and familiar sayings

    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
    345
    +0.13
    0.6%
    70
    +0.12
    0.5%
    598
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    70
    +0.13
    0.04
    598
    +0.12
    0.02
    274
    +0.11
    0.03
    Negative Logits
    ensibly
    -0.85
    ressive
    -0.80
    jri
    -0.79
    ensing
    -0.76
    actions
    -0.75
    ktop
    -0.74
    andestine
    -0.73
    abus
    -0.70
    acting
    -0.70
    medi
    -0.69
    POSITIVE LOGITS
     proverb
    1.04
     motto
    0.90
     aph
    0.82
     mantra
    0.79
     mant
    0.77
     prophes
    0.76
     \"
    0.75
     Blessed
    0.72
     cursed
    0.72
     Thou
    0.71
    Activations Density 0.204%

    No Known Activations