© 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. 3832
    Prev
    Next
    INDEX
    Explanations

    information about races, competitions, and achievements related to sports and events

    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
    346
    +0.12
    0.6%
    94
    +0.09
    0.4%
    343
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    603
    +0.12
    0.04
    24
    +0.09
    0.03
    269
    +0.09
    0.03
    Negative Logits
    soType
    -0.87
     distrust
    -0.83
     Manafort
    -0.80
    Crime
    -0.77
     mistrust
    -0.76
    hate
    -0.76
    FBI
    -0.75
    agency
    -0.73
     intrusive
    -0.72
     informants
    -0.71
    POSITIVE LOGITS
     runner
    1.17
     endurance
    1.08
     runners
    1.06
     solo
    0.99
     medal
    0.98
     paddle
    0.97
     feats
    0.97
     sprint
    0.97
     climb
    0.96
     marathon
    0.96
    Activations Density 0.617%

    No Known Activations