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. 21335
    Prev
    Next
    INDEX
    Explanations

    steps or instructions on how to do a particular task correctly

    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
    366
    +0.12
    0.5%
    497
    +0.11
    0.5%
    459
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    184
    +0.12
    0.03
    105
    +0.11
    0.04
    435
    +0.10
    0.03
    Negative Logits
    Today
    -0.71
    democracy
    -0.65
    humans
    -0.65
     mockery
    -0.64
    Dub
    -0.64
     Canad
    -0.62
     descendants
    -0.62
    john
    -0.62
    Tomorrow
    -0.61
    ogenesis
    -0.61
    POSITIVE LOGITS
     yourself
    1.09
     carefully
    0.92
     BEFORE
    0.92
     beforehand
    0.91
     minimize
    0.89
     preferably
    0.89
     yourselves
    0.89
     spacing
    0.84
     your
    0.84
     Yourself
    0.83
    Activations Density 0.425%

    No Known Activations