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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 26706
    Prev
    Next
    INDEX
    Explanations

    the phrase "come up."

    oai_token-act-pair · gpt-4o-miniTriggered 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
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    125
    +0.11
    0.5%
    530
    +0.10
    0.5%
    739
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    125
    +0.11
    0.02
    322
    +0.10
    0.02
    553
    +0.10
    0.01
    Negative Logits
    Js
    -0.72
     CLSID
    -0.69
    wings
    -0.67
    gee
    -0.65
     Ones
    -0.62
    cise
    -0.61
    cas
    -0.59
    arine
    -0.59
    oly
    -0.58
    ATA
    -0.57
    POSITIVE LOGITS
    river
    0.83
    olicy
    0.75
    votes
    0.75
     GOODMAN
    0.74
    vote
    0.74
     unexpectedly
    0.73
     spontaneously
    0.72
    bid
    0.68
    cloth
    0.68
    adesh
    0.68
    Activations Density 0.019%

    No Known Activations