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

    Neuronpedia

    Jacobian LensNEW
    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. 6103
    Prev
    Next
    INDEX
    Explanations

    questions and feedback prompts related to content or products

    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
    380
    +0.12
    0.6%
    604
    +0.12
    0.5%
    726
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    726
    +0.12
    0.04
    604
    +0.12
    0.03
    380
    +0.10
    0.03
    Negative Logits
    Traditional
    -0.67
    Ö¼
    -0.64
    DEF
    -0.63
    iral
    -0.62
    ilant
    -0.59
    Imm
    -0.58
    virt
    -0.57
    unin
    -0.57
     Auschwitz
    -0.57
    £ı
    -0.57
    POSITIVE LOGITS
     feedback
    0.94
     sugg
    0.94
     suggestions
    0.92
     Feedback
    0.89
     comments
    0.89
    comments
    0.85
     yourself
    0.85
     compliments
    0.78
     Tweet
    0.77
     questions
    0.76
    Activations Density 0.146%

    No Known Activations