© 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 24576
    9. 8-RES_FS24576-JB
    11. 7021
    Prev
    Next
    INDEX
    Explanations

    dialogue exchanges within scripts

    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
    314
    +0.25
    1.3%
    160
    +0.22
    1.1%
    442
    +0.18
    0.9%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    160
    +0.25
    0.11
    314
    +0.22
    0.03
    442
    +0.18
    0.08
    Negative Logits
     Bacon
    -1.00
     Grill
    -0.98
     Smoke
    -0.95
     Pages
    -0.93
     Bag
    -0.93
     Chocolate
    -0.92
     Page
    -0.92
     Nine
    -0.91
     Meter
    -0.91
     Chili
    -0.91
    POSITIVE LOGITS
    subject
    1.53
    last
    1.44
    move
    1.43
    administ
    1.43
    focus
    1.43
    moving
    1.42
    cond
    1.42
    tra
    1.41
    future
    1.40
    nob
    1.39
    Activations Density 0.365%

    No Known Activations