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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 12288
    5. 8-RES_FS12288-JB
    6. 1926
    Prev
    Next
    INDEX
    Explanations

    programming code snippets and technical terms related to programming

    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
    160
    +0.12
    0.5%
    640
    +0.10
    0.5%
    314
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    160
    +0.12
    0.03
    345
    +0.10
    0.02
    701
    +0.10
    0.02
    Negative Logits
     dogged
    -0.78
     plateau
    -0.73
     surpr
    -0.71
     buoy
    -0.71
     brushes
    -0.70
     bitters
    -0.70
     whistle
    -0.69
     dipping
    -0.69
     tire
    -0.67
     moderates
    -0.66
    POSITIVE LOGITS
    SELECT
    1.24
    ./
    1.17
    BuyableInstoreAndOnline
    1.10
    default
    1.08
    Hello
    1.07
    ../
    1.01
    Usage
    0.98
    wcsstore
    0.97
    Untitled
    0.95
    foo
    0.95
    Activations Density 0.068%

    No Known Activations