© 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 49152
    9. 8-RES_FS49152-JB
    11. 25955
    Prev
    Next
    INDEX
    Explanations

    mentions of the term "Af" which likely relates to specific countries or regions in Africa

    oai_token-act-pair · gpt-4o-mini
    Triggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs49152-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    49,152
    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
    288
    +0.15
    0.7%
    311
    +0.13
    0.6%
    187
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    311
    +0.15
    0.01
    187
    +0.13
    0.01
    103
    +0.10
    0.01
    Negative Logits
    soDeliveryDate
    -0.85
     guiActiveUnfocused
    -0.85
    ĸļ
    -0.75
    INESS
    -0.73
    è¯
    -0.71
     MFT
    -0.68
    ORED
    -0.66
     Passion
    -0.65
    assetsadobe
    -0.65
    ãģ®éŃĶ
    -0.65
    POSITIVE LOGITS
    rika
    1.16
    raid
    1.14
    rica
    1.11
    onso
    1.07
    ro
    1.02
    raction
    0.99
    riend
    0.95
    icion
    0.94
    ghan
    0.94
    rican
    0.94
    Activations Density 0.003%

    No Known Activations