© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI 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. 16770
    Prev
    Next
    INDEX
    Explanations

    phrases related to legal actions and conflicts

    oai_token-act-pair · gpt-3.5-turboTriggered 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
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    414
    +0.10
    0.4%
    126
    +0.09
    0.4%
    720
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    705
    +0.10
    0.02
    516
    +0.09
    0.02
    665
    +0.09
    0.02
    Negative Logits
     Gloss
    -0.81
     âĵĺ
    -0.67
     sinks
    -0.66
    terson
    -0.65
     Cake
    -0.64
     rides
    -0.63
     Drops
    -0.63
    olds
    -0.63
     Floating
    -0.62
    arthed
    -0.62
    POSITIVE LOGITS
    respons
    0.88
     instinct
    0.84
     respons
    0.83
     judgement
    0.82
     behalf
    0.80
     conscience
    0.77
     responsibly
    0.77
    ively
    0.77
     accordance
    0.77
     impuls
    0.76
    Activations Density 0.112%

    No Known Activations