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 1536
    5. 8-RES_FS1536-JB
    6. 223
    Prev
    Next
    INDEX
    Explanations

    references to movie-related content, particularly Marvel Cinematic Universe movies, quotes from movie personalities, and information about movie premieres

    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
    477
    +0.11
    0.6%
    515
    +0.10
    0.5%
    610
    +0.09
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    232
    +0.11
    0.13
    284
    +0.10
    0.13
    373
    +0.09
    -0.10
    Negative Logits
    osaurus
    -0.79
    milo
    -0.75
    gow
    -0.67
    xit
    -0.64
    staking
    -0.64
    lav
    -0.64
    elist
    -0.63
    hift
    -0.63
    rss
    -0.62
    cens
    -0.61
    POSITIVE LOGITS
    ĸ
    0.65
    Ń
    0.65
    ±
    0.61
    ļ
    0.59
    ãĥ¼ãĥĨ
    0.57
    µ
    0.56
    ãĤ¨
    0.55
     riches
    0.55
    ³
    0.55
    ī
    0.54
    Activations Density 4.166%

    No Known Activations