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

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 24576
    9. 8-RES_FS24576-JB
    11. 14433
    Prev
    Next
    INDEX
    Explanations

    disclaimers and copyright-related text

    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
    640
    +0.18
    0.9%
    6
    +0.12
    0.5%
    481
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    6
    +0.18
    0.08
    640
    +0.12
    0.07
    679
    +0.10
    0.06
    Negative Logits
     snowball
    -0.75
     crunch
    -0.71
     rallying
    -0.70
     overpowered
    -0.70
     shedding
    -0.69
     induct
    -0.68
     buried
    -0.68
     beating
    -0.67
     amazingly
    -0.67
     cradle
    -0.66
    POSITIVE LOGITS
     Please
    1.48
     Therefore
    1.43
     Unless
    1.34
     Additionally
    1.31
     PLEASE
    1.28
     However
    1.28
     Otherwise
    1.28
     Regardless
    1.24
     Accordingly
    1.24
     Furthermore
    1.23
    Activations Density 0.249%

    No Known Activations