© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI 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. 24390
    Prev
    Next
    INDEX
    Explanations

    phrases or terms that indicate a division or duality, particularly involving the word "half."

    oai_token-act-pair · gpt-4o-miniTriggered 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
    333
    +0.12
    0.6%
    676
    +0.12
    0.5%
    597
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    676
    +0.12
    0.04
    597
    +0.12
    0.03
    190
    +0.10
    0.03
    Negative Logits
    Trend
    -0.78
    issance
    -0.71
    HCR
    -0.70
    anwhile
    -0.69
    lished
    -0.66
    eers
    -0.65
    uable
    -0.64
    EMENT
    -0.63
    è¦
    -0.63
    berus
    -0.62
    POSITIVE LOGITS
    heartedly
    1.15
    hearted
    0.92
    pipe
    0.90
    moon
    0.88
     dozen
    0.84
    wheel
    0.83
    wit
    0.83
    dozen
    0.77
    staff
    0.73
    mast
    0.72
    Activations Density 0.032%

    No Known Activations