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 24576
    5. 8-RES_FS24576-JB
    6. 9076
    Prev
    Next
    INDEX
    Explanations

    descriptions of breakfast food items

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs24576-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    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
    288
    +0.12
    0.5%
    621
    +0.11
    0.5%
    526
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    594
    +0.12
    0.01
    288
    +0.11
    0.01
    326
    +0.10
    0.01
    Negative Logits
    essee
    -0.77
    abases
    -0.74
     Exile
    -0.74
    ne
    -0.72
    rity
    -0.72
     externalToEVAOnly
    -0.72
    idium
    -0.71
    rors
    -0.69
    bilt
    -0.68
    aic
    -0.68
    POSITIVE LOGITS
     cereal
    1.14
     mornings
    1.00
     buffet
    0.92
    urst
    0.92
     cere
    0.91
     breakfast
    0.90
     toast
    0.78
    erences
    0.76
     morning
    0.76
     pancakes
    0.74
    Activations Density 0.022%

    No Known Activations