© 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. 721
    Prev
    Next
    INDEX
    Explanations

    prepositions followed by a place or time

    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
    314
    +0.12
    0.6%
    607
    +0.11
    0.5%
    371
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    371
    +0.12
    0.05
    311
    +0.11
    0.02
    607
    +0.11
    0.04
    Negative Logits
    FTWARE
    -0.68
    aceutical
    -0.67
     Carmen
    -0.59
     kindly
    -0.56
    Berry
    -0.55
    Reviewer
    -0.55
     WRITE
    -0.55
    itivity
    -0.54
    Winged
    -0.54
     benefiting
    -0.53
    POSITIVE LOGITS
     least
    1.63
    onement
    1.33
    roph
    1.20
    rial
    1.03
    yp
    1.03
    hens
    0.97
    rophic
    0.95
    mosp
    0.94
    las
    0.92
    olls
    0.92
    Activations Density 0.110%

    No Known Activations