© 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. 16052
    Prev
    Next
    INDEX
    Explanations

    criminal and legal terms

    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
    314
    +0.13
    0.6%
    280
    +0.13
    0.6%
    694
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    116
    +0.13
    0.04
    280
    +0.13
    0.04
    199
    +0.10
    0.04
    Negative Logits
    isSpecialOrderable
    -0.98
    soDeliveryDate
    -0.95
    DragonMagazine
    -0.89
    izons
    -0.88
    Tokens
    -0.79
    stals
    -0.70
    Companies
    -0.68
     Zuckerberg
    -0.67
     trough
    -0.67
    ellen
    -0.67
    POSITIVE LOGITS
     manslaughter
    1.39
     misconduct
    1.22
     offences
    1.17
     perjury
    1.14
     violations
    1.13
     theft
    1.11
     felony
    1.10
     wrongdoing
    1.10
     misdemeanor
    1.08
     crimes
    1.06
    Activations Density 0.197%

    No Known Activations