© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 12288
    9. 8-RES_FS12288-JB
    11. 782
    Prev
    Next
    INDEX
    Explanations

    technical terms and programming language keywords

    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
    447
    +0.38
    2.3%
    481
    +0.29
    1.7%
    87
    +0.21
    1.3%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.38
    0.16
    481
    +0.29
    0.13
    359
    +0.21
    0.06
    Negative Logits
     inflamm
    -0.45
     Hoover
    -0.45
     MSI
    -0.45
    later
    -0.44
     Tradable
    -0.44
    heid
    -0.43
     gust
    -0.42
    SPONSORED
    -0.42
     Schiff
    -0.42
     Adolf
    -0.42
    POSITIVE LOGITS
    Wiki
    0.69
    MC
    0.49
    pedia
    0.47
    lander
    0.45
    's
    0.45
     manuals
    0.45
    enn
    0.45
    uno
    0.44
    di
    0.43
    Ê
    0.43
    Activations Density 1.829%

    No Known Activations