© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 49152
    9. 8-RES_FS49152-JB
    11. 37186
    Prev
    Next
    INDEX
    Explanations

    punctuation, particularly commas and their frequencies

    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
    481
    +0.12
    0.6%
    322
    +0.10
    0.5%
    210
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    546
    +0.12
    0.04
    66
    +0.10
    0.04
    170
    +0.09
    0.04
    Negative Logits
    ona
    -0.69
    aster
    -0.67
    iph
    -0.67
    ibo
    -0.64
    ulp
    -0.64
    Cause
    -0.63
    anish
    -0.62
    aran
    -0.62
    icut
    -0.62
    adas
    -0.62
    POSITIVE LOGITS
     coupled
    0.89
     chances
    0.81
     culminating
    0.74
     logically
    0.72
     surely
    0.72
     tensions
    0.72
     understandably
    0.71
     uncertainties
    0.69
     speculation
    0.69
     hopefully
    0.69
    Activations Density 0.148%

    No Known Activations