© 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. 14853
    Prev
    Next
    INDEX
    Explanations

    information related to political events and legal cases

    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
    184
    +0.09
    0.4%
    703
    +0.09
    0.4%
    75
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    102
    +0.09
    0.06
    703
    +0.09
    0.04
    55
    +0.09
    0.05
    Negative Logits
    Guide
    -0.73
    Dispatch
    -0.63
    Hat
    -0.62
     refres
    -0.62
    party
    -0.62
    Repeat
    -0.61
    î
    -0.61
    urry
    -0.61
    GI
    -0.60
    raved
    -0.59
    POSITIVE LOGITS
     notoriously
    0.92
     incidentally
    0.80
     famously
    0.79
     notorious
    0.77
     basically
    0.76
     reportedly
    0.74
     essentially
    0.72
    enegger
    0.72
     renamed
    0.72
     comprised
    0.71
    Activations Density 0.397%

    No Known Activations