© 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. 15486
    Prev
    Next
    INDEX
    Explanations

    information related to health, medical conditions, and family health-related issues

    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
    168
    +0.12
    0.5%
    86
    +0.10
    0.5%
    368
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    278
    +0.12
    0.04
    368
    +0.10
    0.04
    86
    +0.10
    0.04
    Negative Logits
     Franch
    -0.73
    Chain
    -0.70
     Graveyard
    -0.68
    Operation
    -0.66
    Installation
    -0.64
    Evil
    -0.64
    bye
    -0.63
    xxxxxxxx
    -0.63
     sr
    -0.61
    pb
    -0.61
    POSITIVE LOGITS
     disproportionately
    1.42
     markedly
    1.14
     statistically
    1.14
     poorer
    1.14
     significantly
    1.14
     outper
    1.13
     overwhelmingly
    1.06
     median
    1.04
     healthier
    0.99
     wealthier
    0.98
    Activations Density 0.342%

    No Known Activations