© 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. 13558
    Prev
    Next
    INDEX
    Explanations

    technical specifications or features of a product, possibly related to sports equipment

    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
    640
    +0.28
    1.3%
    447
    +0.20
    1.0%
    314
    +0.18
    0.9%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    640
    +0.28
    0.09
    160
    +0.20
    0.06
    179
    +0.18
    0.04
    Negative Logits
    hement
    -0.79
     Slater
    -0.73
     Niet
    -0.72
     Instr
    -0.71
     Dollar
    -0.70
     destro
    -0.69
     LI
    -0.66
     Bark
    -0.64
     Daw
    -0.64
     Ont
    -0.64
    POSITIVE LOGITS
    join
    0.96
    hello
    0.87
    condition
    0.85
    ETF
    0.82
    minecraft
    0.80
    natureconservancy
    0.80
    harm
    0.80
    press
    0.79
    gain
    0.78
    Sorry
    0.78
    Activations Density 0.171%

    No Known Activations