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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsBlogSlackPrivacy & 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. 362
    Prev
    Next
    INDEX
    Explanations

    company names and financial information

    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
    373
    +0.17
    0.8%
    496
    +0.11
    0.5%
    299
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    373
    +0.17
    -0.08
    232
    +0.11
    0.16
    659
    +0.11
    0.16
    Negative Logits
     orphans
    -0.59
    ught
    -0.58
    glers
    -0.57
     Ys
    -0.56
     provocation
    -0.54
    java
    -0.51
     Autob
    -0.51
     Generations
    -0.50
     CPR
    -0.49
     CRC
    -0.49
    POSITIVE LOGITS
    nih
    0.95
    ãĤ¼ãĤ¦ãĤ¹
    0.84
    nesota
    0.82
    agement
    0.77
    ciating
    0.76
    thia
    0.75
    atis
    0.73
    azeera
    0.71
    auna
    0.69
    ãĥĬ
    0.69
    Activations Density 4.755%

    No Known Activations