© 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 49152
    9. 8-RES_FS49152-JB
    11. 45612
    Prev
    Next
    INDEX
    Explanations

    references to lightning or related imagery

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs49152-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    49,152
    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
    +0.11
    0.5%
    90
    +0.09
    0.4%
    85
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    90
    +0.11
    0.01
    326
    +0.09
    0.01
    678
    +0.09
    0.01
    Negative Logits
    glers
    -0.96
    ļéĨĴ
    -0.85
    nee
    -0.84
    ablishment
    -0.81
    alse
    -0.81
    eday
    -0.80
    yrinth
    -0.78
    VICE
    -0.78
    rador
    -0.77
    dit
    -0.77
    POSITIVE LOGITS
    bolt
    1.27
     bolt
    1.22
     bolts
    1.17
    storms
    1.01
     strike
    1.01
     strikes
    0.98
     rod
    0.97
     rays
    0.96
     ray
    0.94
    storm
    0.91
    Activations Density 0.004%

    No Known Activations