© 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. 7930
    Prev
    Next
    INDEX
    Explanations

    proper nouns and names, especially those related to individuals

    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
    373
    +0.19
    0.9%
    379
    +0.12
    0.6%
    745
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    373
    +0.19
    -0.01
    232
    +0.12
    0.04
    406
    +0.09
    0.03
    Negative Logits
    FORMATION
    -0.65
    ciating
    -0.57
    士
    -0.56
    ãģ«
    -0.55
     Slater
    -0.55
     Clockwork
    -0.54
    UTION
    -0.53
     Katrina
    -0.53
     HEAD
    -0.53
     veins
    -0.51
    POSITIVE LOGITS
    chel
    1.01
    ech
    0.99
    rag
    0.99
    ey
    0.97
    zsche
    0.97
    rip
    0.96
    etsu
    0.95
    eca
    0.91
    inis
    0.91
    eger
    0.90
    Activations Density 0.072%

    No Known Activations