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. 26316
    Prev
    Next
    INDEX
    Explanations

    punctuation marks that indicate enumeration or lists

    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
    87
    +0.13
    0.6%
    255
    +0.10
    0.5%
    698
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    546
    +0.13
    0.07
    43
    +0.10
    0.06
    297
    +0.10
    0.06
    Negative Logits
    arie
    -0.61
    ropolitan
    -0.61
    yon
    -0.61
    uber
    -0.59
     Moines
    -0.57
    gow
    -0.57
    paren
    -0.56
    orget
    -0.55
    vae
    -0.54
    irs
    -0.54
    POSITIVE LOGITS
     prompting
    1.18
     hence
    1.09
     thus
    1.04
     resulting
    1.03
     respectively
    1.02
     although
    1.02
     albeit
    1.01
     implying
    1.00
     namely
    0.99
     whereas
    0.99
    Activations Density 0.274%

    No Known Activations