Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsBlogSlackPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 3072
    5. 8-RES_FS3072-JB
    6. 2075
    Prev
    Next
    INDEX
    Explanations

    phrases related to government actions and official statements

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs3072-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    3,072
    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
    +0.14
    0.7%
    488
    +0.13
    0.6%
    762
    +0.13
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    146
    +0.14
    0.08
    488
    +0.13
    0.08
    288
    +0.13
    0.07
    Negative Logits
     Width
    -0.71
    CONT
    -0.68
     Candle
    -0.68
     Plate
    -0.66
     Daylight
    -0.66
     Heaven
    -0.65
     Carnival
    -0.65
     Ancest
    -0.65
     Klu
    -0.63
     Velocity
    -0.63
    POSITIVE LOGITS
     agencies
    1.08
     shutdown
    1.08
     bureaucracy
    1.00
     Accountability
    0.97
     coffers
    0.96
     officials
    0.95
    wide
    0.93
     funded
    0.91
     departments
    0.90
     subsidized
    0.90
    Activations Density 0.583%

    No Known Activations