Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 49152
    5. 8-RES_FS49152-JB
    6. 41190
    Prev
    Next
    INDEX
    Explanations

    instances of the word "tuck" and its variations denoting placement or positioning

    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
    598
    +0.10
    0.5%
    159
    +0.09
    0.4%
    729
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    70
    +0.10
    0.01
    251
    +0.09
    0.01
    326
    +0.09
    0.01
    Negative Logits
    FUL
    -0.69
    KT
    -0.67
    ovi
    -0.64
    umer
    -0.63
    é¾į
    -0.62
     Enhancement
    -0.61
     Blessed
    -0.59
    IAL
    -0.58
    ã
    -0.57
    OME
    -0.57
    POSITIVE LOGITS
     snug
    0.91
     away
    0.90
     into
    0.88
    ged
    0.87
    leans
    0.86
    away
    0.86
     neatly
    0.84
    weed
    0.83
     onto
    0.80
    pole
    0.80
    Activations Density 0.012%

    No Known Activations