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. 26159
    Prev
    Next
    INDEX
    Explanations

    mentions of winter and related seasonal themes

    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
    617
    +0.12
    0.6%
    575
    +0.12
    0.6%
    554
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    617
    +0.12
    0.02
    575
    +0.12
    0.02
    571
    +0.11
    0.02
    Negative Logits
    Pages
    -0.79
    aurus
    -0.75
    angler
    -0.73
    ocratic
    -0.72
    raltar
    -0.70
    reme
    -0.70
    Points
    -0.70
    anooga
    -0.69
    reprene
    -0.69
     Unch
    -0.69
    POSITIVE LOGITS
    thur
    0.99
    time
    0.86
     Wonderland
    0.86
     coats
    0.86
     sol
    0.83
     rains
    0.82
     winters
    0.80
     mornings
    0.80
     temperatures
    0.79
    flake
    0.79
    Activations Density 0.013%

    No Known Activations