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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 768
    5. 8-RES_FS768-JB
    6. 530
    Prev
    Next
    INDEX
    Explanations

    scientific and astronomical terms and concepts

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs768-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    768
    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
    +0.13
    0.6%
    570
    +0.11
    0.5%
    668
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    678
    +0.13
    0.15
    570
    +0.11
    0.03
    289
    +0.09
    0.00
    Negative Logits
    street
    -0.89
    pkg
    -0.87
     Homeless
    -0.86
     lawy
    -0.84
     professionalism
    -0.82
     Promotion
    -0.82
    Street
    -0.79
     Criminal
    -0.79
     promo
    -0.79
     Equity
    -0.79
    POSITIVE LOGITS
     electrons
    1.44
     molecules
    1.42
     particles
    1.40
     atoms
    1.36
     photons
    1.35
     emitted
    1.32
     spherical
    1.28
     molecule
    1.27
     nucle
    1.27
     hydrogen
    1.27
    Activations Density 8.256%

    No Known Activations