© 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 3072
    9. 8-RES_FS3072-JB
    11. 772
    Prev
    Next
    INDEX
    Explanations

    quotes attributed to people

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs3072-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    3,072
    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
    209
    +0.10
    0.4%
    121
    +0.10
    0.4%
    373
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    121
    +0.10
    0.06
    332
    +0.10
    0.05
    101
    +0.09
    0.06
    Negative Logits
    mol
    -0.72
    spir
    -0.71
    eneg
    -0.71
    à¼
    -0.70
    enfranch
    -0.69
    à¦
    -0.68
    anuts
    -0.67
     routed
    -0.67
    {"
    -0.63
    Login
    -0.63
    POSITIVE LOGITS
     Diane
    0.94
     analyst
    0.90
     Angela
    0.89
     Donna
    0.87
     Melissa
    0.86
     Dr
    0.86
     Joyce
    0.86
     Sheila
    0.85
     Laura
    0.85
     researcher
    0.84
    Activations Density 0.444%

    No Known Activations