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 768
    5. 8-RES_FS768-JB
    6. 23
    Prev
    Next
    INDEX
    Explanations

    financial and business terms

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs768-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    768
    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
    +0.16
    0.7%
    728
    +0.12
    0.5%
    739
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    393
    +0.16
    0.16
    748
    +0.12
    0.09
    728
    +0.11
    0.09
    Negative Logits
    :(
    -0.71
    actory
    -0.66
    ugi
    -0.66
    isks
    -0.65
    places
    -0.65
    Contents
    -0.65
    bo
    -0.63
    &
    -0.62
    >[
    -0.62
    stocks
    -0.62
    POSITIVE LOGITS
     others
    1.18
     Sons
    1.04
     Associates
    1.02
     other
    1.01
    ERSON
    0.91
    rogens
    0.89
     assorted
    0.87
     Others
    0.84
     vice
    0.83
     consequently
    0.81
    Activations Density 2.544%

    No Known Activations