© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW 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. 15002
    Prev
    Next
    INDEX
    Explanations

    references to sports scores and leads

    oai_token-act-pair · gpt-3.5-turbo
    Triggered 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
    Neuron Alignment
    Index
    Value
    % of L₁
    214
    +0.09
    0.4%
    93
    +0.09
    0.4%
    442
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    214
    +0.09
    0.01
    391
    +0.09
    0.01
    259
    +0.09
    0.01
    Negative Logits
     concess
    -0.79
    ROR
    -0.65
    ramid
    -0.64
    irlf
    -0.63
    anism
    -0.62
    entric
    -0.61
     representations
    -0.60
    wagen
    -0.59
    ileaks
    -0.58
     Rebell
    -0.57
    POSITIVE LOGITS
    hunter
    0.79
    breaker
    0.75
    now
    0.73
    drawn
    0.72
    ership
    0.71
     margin
    0.70
    leader
    0.70
    holder
    0.70
    manship
    0.69
    better
    0.68
    Activations Density 0.029%

    No Known Activations