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 49152
    5. 8-RES_FS49152-JB
    6. 23317
    Prev
    Next
    INDEX
    Explanations

    expressions of gratitude, particularly in the context of subscriptions or acknowledgments

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs49152-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    49,152
    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
    240
    +0.11
    0.5%
    310
    +0.10
    0.5%
    604
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    604
    +0.11
    0.02
    131
    +0.10
    0.01
    359
    +0.10
    0.01
    Negative Logits
    panic
    -0.60
    ERROR
    -0.59
    tein
    -0.58
    aiman
    -0.57
     Worse
    -0.56
    chio
    -0.56
     Mig
    -0.55
    orer
    -0.55
    xit
    -0.55
    OSP
    -0.55
    POSITIVE LOGITS
     subscribing
    0.80
    DragonMagazine
    0.67
     dues
    0.65
    scribe
    0.62
    à¤
    0.59
     à¤
    0.59
    unity
    0.58
    à¥
    0.56
     accordingly
    0.55
     generously
    0.55
    Activations Density 0.009%

    No Known Activations