© 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 3072
    9. 8-RES_FS3072-JB
    11. 684
    Prev
    Next
    INDEX
    Explanations

    the word "when" at the beginning of a sentence

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs3072-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    3,072
    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
    706
    +0.11
    0.5%
    476
    +0.11
    0.5%
    393
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    582
    +0.11
    0.08
    30
    +0.11
    0.07
    26
    +0.10
    0.08
    Negative Logits
    agin
    -0.80
    zzi
    -0.67
    aido
    -0.65
    bear
    -0.64
    whatever
    -0.64
    rolet
    -0.62
    edly
    -0.61
    him
    -0.61
    kaya
    -0.61
    stead
    -0.61
    POSITIVE LOGITS
    soever
    1.38
     confronted
    1.12
     faced
    1.02
     asked
    0.96
     pressed
    0.95
     viewed
    0.87
     encountering
    0.87
     compared
    0.86
     comparing
    0.82
     contrasted
    0.81
    Activations Density 0.434%

    No Known Activations