© 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. 3286
    Prev
    Next
    INDEX
    Explanations

    instances where a word "an" is followed by a noun

    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
    572
    +0.12
    0.6%
    393
    +0.12
    0.6%
    523
    +0.12
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    572
    +0.12
    0.09
    362
    +0.12
    0.08
    705
    +0.12
    0.08
    Negative Logits
     branches
    -0.70
    birds
    -0.68
    stones
    -0.67
     pupils
    -0.66
     bills
    -0.65
     nests
    -0.65
    Topics
    -0.64
    ments
    -0.64
    BLE
    -0.64
    hood
    -0.63
    POSITIVE LOGITS
    abolic
    1.21
    cients
    1.09
    esthetic
    1.00
    ointed
    0.97
    agram
    0.95
     enormous
    0.94
    onym
    0.94
    omaly
    0.94
    alyses
    0.93
    aer
    0.93
    Activations Density 0.251%

    No Known Activations