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

    Neuronpedia

    APISteerSAE EvalsBlog/PodcastNEWSlackPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlog/PodcastGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 3072
    5. 8-RES_FS3072-JB
    6. 1577
    Prev
    Next
    INDEX
    Explanations

    sports-related terms and actions, particularly in the context of fighting sports like UFC

    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
    268
    +0.10
    0.5%
    590
    +0.09
    0.4%
    452
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.10
    0.07
    652
    +0.09
    0.05
    326
    +0.09
    0.06
    Negative Logits
    Newsletter
    -0.81
    aceae
    -0.80
    ridges
    -0.77
    ulhu
    -0.76
     Gutenberg
    -0.76
    minecraft
    -0.75
     Helpful
    -0.74
    landers
    -0.73
    quist
    -0.73
    sonian
    -0.70
    POSITIVE LOGITS
     MMA
    1.13
     bout
    1.13
     rematch
    1.08
     boxing
    1.08
     boxer
    1.03
     wrestlers
    1.00
     bouts
    0.98
     Boxing
    0.96
     UFC
    0.95
    UFC
    0.95
    Activations Density 1.021%

    No Known Activations