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 24576
    5. 8-RES_FS24576-JB
    6. 21737
    Prev
    Next
    INDEX
    Explanations

    mentions of movies, TV shows, books, or other forms of media

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs24576-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    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
    48
    +0.11
    0.5%
    567
    +0.09
    0.4%
    517
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    765
    +0.11
    0.07
    43
    +0.09
    0.06
    297
    +0.09
    0.05
    Negative Logits
     WATCHED
    -0.70
    raq
    -0.69
    ħĭ
    -0.68
    ase
    -0.67
    pec
    -0.64
    irie
    -0.64
    omore
    -0.64
    oir
    -0.63
    ggy
    -0.62
    lash
    -0.62
    POSITIVE LOGITS
     although
    1.00
     aka
    0.96
     respectively
    0.95
     whereas
    0.95
     consisting
    0.95
     meaning
    0.93
     wherein
    0.92
     which
    0.90
     featuring
    0.89
     headquartered
    0.85
    Activations Density 0.277%

    No Known Activations