© 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. 769
    Prev
    Next
    INDEX
    Explanations

    the word "swing" followed by numbers or hyphens

    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
    288
    +0.13
    0.6%
    240
    +0.10
    0.5%
    143
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.13
    0.02
    326
    +0.10
    0.01
    374
    +0.09
    0.01
    Negative Logits
    xia
    -0.71
    sis
    -0.70
     Seym
    -0.69
     ABE
    -0.69
    src
    -0.68
    士
    -0.67
     Lauder
    -0.65
    HCR
    -0.65
    resp
    -0.63
    udes
    -0.62
    POSITIVE LOGITS
     swinging
    0.86
    mill
    0.84
     swing
    0.83
     pend
    0.81
    erer
    0.79
    shape
    0.79
    grass
    0.78
     swings
    0.77
    erman
    0.77
    yip
    0.75
    Activations Density 0.059%

    No Known Activations