© 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 EvalsExportsinterp-engineNEWAPI 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. 4419
    Prev
    Next
    INDEX
    Explanations

    references to lifestyle choices, health, skills development, careers, community health, and greenhouse gases

    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
    163
    +0.13
    0.6%
    288
    +0.11
    0.5%
    19
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.13
    0.10
    163
    +0.11
    0.06
    365
    +0.10
    0.03
    Negative Logits
     Empty
    -0.73
    ĸļ
    -0.73
     Originally
    -0.73
    erity
    -0.70
     Quote
    -0.69
     Sturgeon
    -0.67
     Yue
    -0.66
     Patton
    -0.65
    otto
    -0.64
     Beware
    -0.64
    POSITIVE LOGITS
     categories
    1.37
     scenarios
    1.33
     configurations
    1.30
     contexts
    1.28
     types
    1.28
     regimes
    1.25
     formats
    1.24
     disciplines
    1.23
     situations
    1.23
     fronts
    1.18
    Activations Density 0.536%

    No Known Activations