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. 12125
    Prev
    Next
    INDEX
    Explanations

    conjunctions, particularly the word "and"

    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
    393
    +0.10
    0.5%
    659
    +0.10
    0.4%
    651
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    659
    +0.10
    0.03
    766
    +0.10
    0.03
    728
    +0.09
    0.02
    Negative Logits
    wcs
    -0.84
    illion
    -0.82
    artifacts
    -0.74
    Interstitial
    -0.71
    Contents
    -0.70
    bucks
    -0.70
    capacity
    -0.70
    interstitial
    -0.69
    establishment
    -0.69
    ãĥİ
    -0.68
    POSITIVE LOGITS
     colleagues
    1.42
     others
    1.15
     Associates
    1.14
     colleague
    1.11
     collaborators
    1.03
     coworkers
    0.97
     fellow
    0.93
     researcher
    0.88
     PhD
    0.82
     collaborator
    0.81
    Activations Density 0.062%

    No Known Activations