© 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. 10097
    Prev
    Next
    INDEX
    Explanations

    names of individuals, particularly Paul

    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
    346
    +0.11
    0.5%
    314
    +0.10
    0.5%
    519
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    346
    +0.11
    0.04
    519
    +0.10
    0.03
    117
    +0.10
    0.03
    Negative Logits
    cyl
    -0.72
     lodge
    -0.67
    engineering
    -0.64
    Operation
    -0.63
    SPONSORED
    -0.60
    VICE
    -0.57
    dayName
    -0.56
    Limit
    -0.56
    ../
    -0.56
     tolerant
    -0.56
    POSITIVE LOGITS
    son
    1.12
     McCartney
    1.12
    ine
    1.02
    sen
    1.00
    sonian
    0.97
    ozo
    0.97
     Krugman
    0.97
    sburg
    0.93
    ieu
    0.90
     Manafort
    0.89
    Activations Density 0.066%

    No Known Activations