© 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 1536
    9. 8-RES_FS1536-JB
    11. 1131
    Prev
    Next
    INDEX
    Explanations

    physics-related terms and scientific discoveries

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs1536-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    1,536
    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
    678
    +0.10
    0.5%
    263
    +0.10
    0.4%
    535
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    678
    +0.10
    0.10
    570
    +0.10
    0.01
    647
    +0.09
    0.08
    Negative Logits
    street
    -0.91
    pkg
    -0.89
    iquette
    -0.88
     Homeless
    -0.87
    Veter
    -0.87
     Veteran
    -0.84
    Street
    -0.84
    Campaign
    -0.81
     professionalism
    -0.80
     Casual
    -0.79
    POSITIVE LOGITS
     atoms
    1.39
     molecules
    1.37
     electrons
    1.35
     photons
    1.35
     physicists
    1.32
     orbiting
    1.30
     molecule
    1.30
     gravitational
    1.29
     neutron
    1.27
     particles
    1.26
    Activations Density 4.490%

    No Known Activations