© 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 24576
    9. 8-RES_FS24576-JB
    11. 18837
    Prev
    Next
    INDEX
    Explanations

    the word "step" followed by a numeric value (eg. step 9)

    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
    721
    +0.11
    0.5%
    472
    +0.10
    0.5%
    359
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    754
    +0.11
    0.02
    472
    +0.10
    0.02
    441
    +0.10
    0.02
    Negative Logits
    yip
    -0.71
    selage
    -0.66
    è¦ļéĨĴ
    -0.66
    ores
    -0.65
    oros
    -0.64
    ciating
    -0.64
    ecause
    -0.64
    raid
    -0.63
    ãĥīãĥ©ãĤ´ãĥ³
    -0.62
     Moroc
    -0.62
    POSITIVE LOGITS
     aside
    0.94
     forth
    0.94
    frog
    0.91
     ashore
    0.91
     up
    0.91
    up
    0.84
     forward
    0.84
     toe
    0.82
     foot
    0.82
     onstage
    0.81
    Activations Density 0.022%

    No Known Activations