© 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. 25986
    Prev
    Next
    INDEX
    Explanations

    the word "up" in various contexts

    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
    314
    +0.13
    0.6%
    99
    +0.12
    0.6%
    159
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    533
    +0.13
    0.03
    99
    +0.12
    0.02
    262
    +0.11
    0.02
    Negative Logits
    VICE
    -0.63
    gemony
    -0.62
     understatement
    -0.62
     CLSID
    -0.59
     separat
    -0.58
    9999
    -0.56
     Kahn
    -0.55
     Xin
    -0.55
     Lumpur
    -0.54
    hiba
    -0.54
    POSITIVE LOGITS
    raised
    1.07
    stairs
    0.96
    grading
    0.94
    dates
    0.91
    rated
    0.87
    ended
    0.87
    graded
    0.86
    dating
    0.85
    swing
    0.85
    grades
    0.83
    Activations Density 0.021%

    No Known Activations