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 1536
    5. 8-RES_FS1536-JB
    6. 753
    Prev
    Next
    INDEX
    Explanations

    phrases related to beginnings or starting points

    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
    481
    +0.10
    0.4%
    298
    +0.10
    0.4%
    762
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    669
    +0.10
    0.09
    608
    +0.10
    0.12
    86
    +0.09
    0.08
    Negative Logits
    iste
    -0.65
    soDeliveryDate
    -0.63
    lass
    -0.61
    tec
    -0.59
    izes
    -0.58
    redits
    -0.57
     Courier
    -0.57
    arton
    -0.57
    rones
    -0.57
    ize
    -0.56
    POSITIVE LOGITS
     putting
    1.52
     delivering
    1.51
     getting
    1.50
     creating
    1.50
     completing
    1.48
     releasing
    1.47
     gaining
    1.44
     becoming
    1.44
     having
    1.44
     maintaining
    1.43
    Activations Density 3.443%

    No Known Activations