© 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. 14877
    Prev
    Next
    INDEX
    Explanations

    words related to societal issues and controversies

    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
    637
    +0.12
    0.6%
    257
    +0.10
    0.5%
    315
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    637
    +0.12
    0.02
    194
    +0.10
    0.02
    257
    +0.10
    0.02
    Negative Logits
    abase
    -0.84
    icion
    -0.74
    gone
    -0.74
    aimon
    -0.66
    gem
    -0.61
    ignt
    -0.60
    aiman
    -0.59
    ibles
    -0.59
    oplan
    -0.56
    akuya
    -0.56
    POSITIVE LOGITS
    Article
    0.77
     ARTICLE
    0.72
     VIDEOS
    0.67
     reprinted
    0.66
     »
    0.66
    Subscribe
    0.63
     BELOW
    0.62
    ¢
    0.62
     Loading
    0.60
     Travels
    0.60
    Activations Density 0.013%

    No Known Activations