© 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
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 24576
    5. 8-RES_FS24576-JB
    6. 7313
    Prev
    Next
    INDEX
    Explanations

    specific terms indicating a combination, connection, or association between different elements

    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
    514
    +0.12
    0.5%
    403
    +0.10
    0.5%
    551
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    665
    +0.12
    0.04
    332
    +0.10
    0.03
    125
    +0.10
    0.03
    Negative Logits
    girls
    -0.72
    AH
    -0.68
    ahs
    -0.66
    bern
    -0.64
    worms
    -0.61
    sburg
    -0.61
     Semin
    -0.61
    boys
    -0.61
    dark
    -0.61
    udeb
    -0.59
    POSITIVE LOGITS
    isons
    0.90
     oppos
    0.85
     thereto
    0.80
     logically
    0.78
     nicely
    0.77
     seamlessly
    0.76
     neatly
    0.76
    icut
    0.76
     synerg
    0.75
    ison
    0.75
    Activations Density 0.124%

    No Known Activations