© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 12288
    9. 8-RES_FS12288-JB
    11. 9727
    Prev
    Next
    INDEX
    Explanations

    names of actors or characters in movies and TV shows

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs12288-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    12,288
    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
    373
    +0.13
    0.6%
    615
    +0.11
    0.5%
    572
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    572
    +0.13
    0.04
    582
    +0.11
    0.03
    499
    +0.11
    0.03
    Negative Logits
     newsletters
    -1.06
    wcs
    -0.83
    Statistics
    -0.82
    Spread
    -0.77
    ickr
    -0.74
    SOURCE
    -0.71
     newsletter
    -0.71
     "$:/
    -0.70
    foundation
    -0.70
    £ı
    -0.69
    POSITIVE LOGITS
     villain
    0.94
     actor
    0.89
     protagonist
    0.88
     Dracula
    0.88
     narrator
    0.87
     actress
    0.87
     heroine
    0.86
     antagonist
    0.86
     characters
    0.86
     Characters
    0.85
    Activations Density 0.252%

    No Known Activations