© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 12288
    9. 8-RES_FS12288-JB
    11. 1868
    Prev
    Next
    INDEX
    Explanations

    differences in physical attributes or features between similar objects

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs12288-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    12,288
    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
    362
    +0.24
    1.1%
    393
    +0.22
    1.0%
    608
    +0.12
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    362
    +0.24
    0.10
    608
    +0.22
    0.07
    393
    +0.12
    0.06
    Negative Logits
     followed
    -0.76
     besides
    -0.75
     ahead
    -0.74
     alike
    -0.73
     ********************************
    -0.71
    won
    -0.71
     Became
    -0.71
    ãĥ»
    -0.69
    amar
    -0.68
     reminded
    -0.68
    POSITIVE LOGITS
     usual
    1.30
     rest
    1.25
     aforementioned
    1.17
     proverbial
    1.10
     slightest
    1.10
     ones
    1.06
     entirety
    1.06
     original
    1.05
     nearest
    1.03
     originals
    1.01
    Activations Density 0.363%

    No Known Activations