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 768
    5. 8-RES_FS768-JB
    6. 566
    Prev
    Next
    INDEX
    Explanations

    introductory sentences or sections

    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
    366
    +0.13
    0.6%
    339
    +0.12
    0.6%
    290
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    366
    +0.13
    0.11
    339
    +0.12
    0.11
    162
    +0.11
    0.09
    Negative Logits
    tics
    -0.71
     Canaver
    -0.71
    sung
    -0.67
     Gould
    -0.63
    vor
    -0.61
    roller
    -0.60
    sav
    -0.60
    rolet
    -0.59
    aths
    -0.59
    cling
    -0.59
    POSITIVE LOGITS
     responders
    1.33
     baseman
    1.18
     impressions
    0.96
     glance
    0.95
     blush
    0.93
     foray
    0.89
    born
    0.84
     impression
    0.81
     lady
    0.81
     glimpse
    0.80
    Activations Density 1.924%

    No Known Activations