© 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. 14456
    Prev
    Next
    INDEX
    Explanations

    commands and procedures in a technical setup or tutorial

    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
    373
    +0.16
    0.7%
    745
    +0.14
    0.6%
    545
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    326
    +0.16
    0.02
    680
    +0.14
    0.02
    640
    +0.10
    0.03
    Negative Logits
    hemer
    -0.76
    iquid
    -0.70
     Coffin
    -0.68
     Kahn
    -0.66
     Barron
    -0.64
     drawings
    -0.64
    ricular
    -0.63
     Winged
    -0.63
     wrinkles
    -0.61
     boobs
    -0.61
    POSITIVE LOGITS
    sudo
    1.35
     sudo
    1.23
    cmd
    1.08
     npm
    1.02
     ./
    1.02
     grep
    0.99
    pkg
    0.98
    command
    0.93
    exe
    0.93
    git
    0.93
    Activations Density 0.092%

    No Known Activations