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 24576
    5. 8-RES_FS24576-JB
    6. 23645
    Prev
    Next
    INDEX
    Explanations

    punctuation marks and sentence boundaries

    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
    679
    +0.16
    0.8%
    640
    +0.15
    0.7%
    464
    +0.13
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    6
    +0.16
    0.14
    679
    +0.15
    0.12
    430
    +0.13
    0.11
    Negative Logits
    kefeller
    -0.83
     democrat
    -0.74
     sche
    -0.73
     challeng
    -0.72
     sustainable
    -0.71
     diseng
    -0.71
     insur
    -0.71
     affili
    -0.70
     prey
    -0.66
     conscience
    -0.66
    POSITIVE LOGITS
     Initially
    1.17
     Its
    1.14
     However
    1.13
     Previously
    1.08
     Normally
    1.05
     It
    1.04
     Firstly
    1.03
     Each
    1.03
     Although
    1.03
     Though
    1.03
    Activations Density 0.549%

    No Known Activations