© 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. 15014
    Prev
    Next
    INDEX
    Explanations

    mentions of food contamination or health hazards related to food products

    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
    447
    +0.14
    0.6%
    314
    +0.12
    0.5%
    253
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    607
    +0.14
    0.01
    54
    +0.12
    0.01
    6
    +0.11
    0.01
    Negative Logits
    ously
    -0.75
     Krug
    -0.68
     Galile
    -0.68
    uers
    -0.67
     slope
    -0.66
    ãĥ¼ãĥĨãĤ£
    -0.65
     slopes
    -0.64
    xes
    -0.64
    govtrack
    -0.63
    olas
    -0.63
    POSITIVE LOGITS
    etc
    1.09
    please
    0.90
    walking
    0.86
    where
    0.85
    err
    0.83
    right
    0.82
    especially
    0.81
    which
    0.79
    MORE
    0.78
    BUT
    0.78
    Activations Density 0.010%

    No Known Activations