© 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. 2955
    Prev
    Next
    INDEX
    Explanations

    parenthetical information such as dates and ratings

    oai_token-act-pair · gpt-3.5-turboTriggered 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
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    314
    +0.21
    1.0%
    640
    +0.18
    0.8%
    160
    +0.17
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    160
    +0.21
    0.09
    314
    +0.18
    0.03
    627
    +0.17
    0.05
    Negative Logits
     intelligent
    -0.79
     retard
    -0.76
     everyday
    -0.76
     occupancy
    -0.74
     sharp
    -0.73
     instinct
    -0.73
     mature
    -0.72
     intensive
    -0.72
     flow
    -0.72
     lull
    -0.71
    POSITIVE LOGITS
    which
    1.56
    pictured
    1.54
    see
    1.53
    although
    1.48
    formerly
    1.45
    sic
    1.40
    shown
    1.37
    including
    1.36
    along
    1.35
    emphasis
    1.34
    Activations Density 0.139%

    No Known Activations