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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & 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. 25975
    Prev
    Next
    INDEX
    Explanations

    references to disclaimer sections or indications to continue reading visually out of context

    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
    373
    +0.11
    0.6%
    530
    +0.09
    0.5%
    348
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    530
    +0.11
    0.01
    683
    +0.09
    0.01
    613
    +0.09
    0.01
    Negative Logits
    venge
    -0.62
    nar
    -0.61
    reb
    -0.59
     mart
    -0.58
    jab
    -0.58
    urion
    -0.58
    owe
    -0.57
    john
    -0.57
     appar
    -0.56
    folk
    -0.56
    POSITIVE LOGITS
     Loading
    0.85
     Below
    0.81
     Thumbnails
    0.74
     Transcript
    0.73
     Advertisement
    0.69
     Continue
    0.68
     Expand
    0.65
     Story
    0.63
     Average
    0.62
     Skip
    0.61
    Activations Density 0.003%

    No Known Activations