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. 25606
    Prev
    Next
    INDEX
    Explanations

    instances of the word "think" and its variations, capturing expressions of thought or consideration

    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
    373
    +0.14
    0.6%
    493
    +0.12
    0.6%
    679
    +0.12
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    287
    +0.14
    0.02
    70
    +0.12
    0.03
    675
    +0.12
    0.02
    Negative Logits
    iations
    -0.69
    kw
    -0.67
    Footnote
    -0.65
    recorded
    -0.64
    versions
    -0.62
    ×Ļ×
    -0.62
    iona
    -0.59
    clad
    -0.59
    SPONSORED
    -0.58
    atile
    -0.58
    POSITIVE LOGITS
     bout
    0.74
     about
    0.73
    stock
    0.72
    enery
    0.70
    pad
    0.68
     ABOUT
    0.67
     Bout
    0.63
    OTAL
    0.61
    asio
    0.61
    noon
    0.60
    Activations Density 0.032%

    No Known Activations