© 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. 3331
    Prev
    Next
    INDEX
    Explanations

    words related to criminal activities and legal terms

    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
    470
    +0.10
    0.5%
    336
    +0.10
    0.4%
    620
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    470
    +0.10
    0.02
    163
    +0.10
    0.03
    232
    +0.09
    0.03
    Negative Logits
    eers
    -0.71
    creen
    -0.71
     cov
    -0.70
    e
    -0.67
    eq
    -0.65
    eas
    -0.63
    eer
    -0.63
    Ĥ¬
    -0.62
     EntityItem
    -0.61
    Ŀ
    -0.61
    POSITIVE LOGITS
    abbit
    1.27
    acing
    1.20
    ussia
    1.20
    angers
    1.15
    ifle
    1.13
    uder
    1.13
    outine
    1.10
    aceutical
    1.09
    agnar
    1.08
    ific
    1.07
    Activations Density 0.043%

    No Known Activations