© 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. 13417
    Prev
    Next
    INDEX
    Explanations

    mentions of political and social issues, including incidents related to violence, law enforcement, and investigations

    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
    1.4%
    314
    +0.13
    0.6%
    230
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    640
    +0.28
    0.18
    94
    +0.13
    0.14
    64
    +0.10
    -0.04
    Negative Logits
     equip
    -0.63
     multic
    -0.61
    xual
    -0.61
    boro
    -0.61
     Trog
    -0.60
     dime
    -0.57
    utical
    -0.57
     desks
    -0.57
     fatig
    -0.56
     innocence
    -0.56
    POSITIVE LOGITS
    ³³³³³³³³
    1.10
    SPONSORED
    1.09
    ³³³³
    1.04
    Scroll
    0.99
    ³³³
    0.98
    ³³³³³³³³³³³³³³³³
    0.97
    However
    0.95
    ccording
    0.86
    PHOTOS
    0.86
    Yet
    0.86
    Activations Density 0.513%

    No Known Activations