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. 23472
    Prev
    Next
    INDEX
    Explanations

    phrases indicating the potential or ability to perform actions

    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
    665
    +0.15
    0.7%
    551
    +0.14
    0.7%
    718
    +0.13
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    49
    +0.15
    0.06
    665
    +0.14
    0.06
    551
    +0.13
    0.05
    Negative Logits
     Coff
    -0.73
     Respons
    -0.71
    ELD
    -0.70
    soDeliveryDate
    -0.65
     Rising
    -0.65
    prem
    -0.63
    famous
    -0.63
    listed
    -0.62
    mails
    -0.61
     Born
    -0.61
    POSITIVE LOGITS
     withstand
    1.26
     afford
    1.22
     cope
    1.20
     convince
    1.15
     overcome
    1.14
     accommodate
    1.13
     communicate
    1.12
     manipulate
    1.12
     muster
    1.11
     achieve
    1.10
    Activations Density 0.088%

    No Known Activations