© 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. 718
    Prev
    Next
    INDEX
    Explanations

    words related to politics, peace, development, and community initiatives

    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
    575
    +0.12
    0.5%
    447
    +0.11
    0.5%
    132
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    575
    +0.12
    0.09
    225
    +0.11
    0.08
    724
    +0.11
    0.08
    Negative Logits
     bluff
    -0.70
    Ly
    -0.62
     previews
    -0.62
     toes
    -0.62
     retract
    -0.60
     sheer
    -0.60
     tacit
    -0.59
     underest
    -0.59
    OUT
    -0.58
     overest
    -0.57
    POSITIVE LOGITS
    cellence
    1.00
     amidst
    0.87
     prevail
    0.81
     wherever
    0.79
     amongst
    0.78
    .","
    0.78
    riors
    0.76
    equality
    0.74
    .</
    0.73
     whilst
    0.72
    Activations Density 0.517%

    No Known Activations