© 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
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 12288
    5. 8-RES_FS12288-JB
    6. 604
    Prev
    Next
    INDEX
    Explanations

    instances where the word "climber" or variations of the word are mentioned

    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
    218
    +0.17
    0.8%
    497
    +0.12
    0.6%
    97
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    588
    +0.17
    0.04
    326
    +0.12
    0.04
    232
    +0.10
    0.04
    Negative Logits
     Hipp
    -0.57
    essel
    -0.57
    ishly
    -0.56
     ISO
    -0.56
    yne
    -0.56
     manifests
    -0.56
     Moose
    -0.55
    ive
    -0.55
     Seraph
    -0.55
    Ju
    -0.54
    POSITIVE LOGITS
    atically
    1.27
    atic
    1.27
    pse
    1.22
    bour
    1.22
    pton
    1.21
    ming
    1.17
    matically
    1.11
    bler
    1.11
    mer
    1.10
    atics
    1.08
    Activations Density 0.174%

    No Known Activations