© 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. 14879
    Prev
    Next
    INDEX
    Explanations

    references to specific groups of people (e.g., "you may have," "you who") in text

    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
    763
    +0.13
    0.6%
    1
    +0.11
    0.5%
    341
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    763
    +0.13
    0.02
    341
    +0.11
    0.02
    49
    +0.09
    0.02
    Negative Logits
    orp
    -0.75
     Canaver
    -0.68
    ipal
    -0.67
    apo
    -0.67
     Shutterstock
    -0.63
     Heist
    -0.62
     Cure
    -0.61
    Layer
    -0.60
     Hick
    -0.60
     srfAttach
    -0.60
    POSITIVE LOGITS
     guys
    1.21
    tub
    1.17
     yourselves
    1.02
     gentlemen
    0.95
     readers
    0.87
     boys
    0.84
     youngsters
    0.81
     know
    0.79
     listeners
    0.79
     remember
    0.78
    Activations Density 0.033%

    No Known Activations