© 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. 14912
    Prev
    Next
    INDEX
    Explanations

    dates and proper nouns related to politics or current events

    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
    739
    +0.13
    0.6%
    530
    +0.13
    0.6%
    373
    +0.12
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    373
    +0.13
    -0.01
    530
    +0.13
    0.05
    406
    +0.12
    0.05
    Negative Logits
    jet
    -0.63
    thouse
    -0.59
    hex
    -0.59
    boys
    -0.58
    leigh
    -0.58
     Britann
    -0.57
    zac
    -0.57
     gag
    -0.56
    ks
    -0.56
    gate
    -0.56
    POSITIVE LOGITS
    IZ
    1.45
    ING
    1.45
    ER
    1.44
    ATE
    1.43
    ITY
    1.41
    NESS
    1.41
    ITIES
    1.40
    URA
    1.40
    IES
    1.40
    ATER
    1.40
    Activations Density 0.120%

    No Known Activations