© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 19990
    Prev
    Next
    INDEX
    Explanations

    references to institutions or organizations

    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
    712
    +0.11
    0.5%
    612
    +0.09
    0.4%
    53
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    53
    +0.11
    0.04
    81
    +0.09
    0.03
    0
    +0.09
    0.03
    Negative Logits
    netflix
    -0.90
    blank
    -0.76
    pedia
    -0.71
    minus
    -0.67
    oidal
    -0.65
     flush
    -0.65
     oneself
    -0.64
    pause
    -0.64
    nation
    -0.63
     droid
    -0.63
    POSITIVE LOGITS
     Corrections
    0.93
     Technology
    0.92
     Telecommunications
    0.92
     Commerce
    0.92
     Health
    0.91
     Ethics
    0.89
     Excellence
    0.89
     Agriculture
    0.88
     Medicine
    0.86
     Standards
    0.86
    Activations Density 0.063%

    No Known Activations