© 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
    3. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 24576
    9. 8-RES_FS24576-JB
    11. 22101
    Prev
    Next
    INDEX
    Explanations

    occurrences of the word "require" and its variations in legislative contexts

    oai_token-act-pair · gpt-4o-miniTriggered 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
    520
    +0.12
    0.5%
    671
    +0.10
    0.5%
    385
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    520
    +0.12
    0.03
    671
    +0.10
    0.02
    713
    +0.09
    0.02
    Negative Logits
    vironment
    -0.78
    tr
    -0.74
    rou
    -0.73
    bench
    -0.73
    arse
    -0.70
    fal
    -0.69
    itsch
    -0.68
    apest
    -0.67
    sm
    -0.66
    mania
    -0.66
    POSITIVE LOGITS
     compliance
    1.15
     applicants
    1.06
     employers
    0.98
     recipients
    0.93
     automakers
    0.90
     adherence
    0.88
     manufacturers
    0.87
     unanim
    0.85
     ISPs
    0.85
     entities
    0.84
    Activations Density 0.045%

    No Known Activations