© 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. 15952
    Prev
    Next
    INDEX
    Explanations

    phrases related to legal matters and committees

    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
    92
    +0.10
    0.4%
    481
    +0.09
    0.4%
    218
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    197
    +0.10
    0.02
    631
    +0.09
    0.02
    287
    +0.09
    0.02
    Negative Logits
     scoop
    -0.80
     siph
    -0.76
     buckets
    -0.72
     finger
    -0.69
     crate
    -0.68
     jars
    -0.67
     pump
    -0.66
     eater
    -0.66
     glances
    -0.65
     toe
    -0.64
    POSITIVE LOGITS
     Aging
    1.20
     behalf
    0.99
     Excellence
    0.92
     Principles
    0.87
     Standards
    0.86
     Settlement
    0.86
     Implementation
    0.85
     Enh
    0.84
     Drugs
    0.83
     Terrorism
    0.82
    Activations Density 0.037%

    No Known Activations