Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsBlogSlackPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 1536
    5. 8-RES_FS1536-JB
    6. 892
    Prev
    Next
    INDEX
    Explanations

    numerical data such as statistics, measurements, and percentages

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs1536-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    1,536
    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
    745
    +0.12
    0.6%
    328
    +0.10
    0.5%
    131
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    495
    +0.12
    0.08
    432
    +0.10
    0.07
    123
    +0.10
    0.07
    Negative Logits
    raltar
    -0.74
    redo
    -0.72
    adobe
    -0.68
    linger
    -0.65
    recy
    -0.65
    rament
    -0.64
    pocket
    -0.61
    heads
    -0.61
    ilial
    -0.60
    ships
    -0.59
    POSITIVE LOGITS
    GHz
    1.05
     billion
    0.94
     million
    0.94
     GHz
    0.91
     GPA
    0.87
     ±
    0.86
     lakh
    0.86
     trillion
    0.84
     crore
    0.84
    3333
    0.81
    Activations Density 0.883%

    No Known Activations