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. 43236
    Prev
    Next
    INDEX
    Explanations

    references to authority, responsibility, or roles involving oversight and enforcement

    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
    562
    +0.12
    0.5%
    756
    +0.10
    0.5%
    531
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    562
    +0.12
    0.01
    665
    +0.10
    0.01
    587
    +0.09
    0.01
    Negative Logits
    Pac
    -0.75
    DragonMagazine
    -0.65
     istg
    -0.64
    laughter
    -0.64
     misses
    -0.64
    ABC
    -0.63
     Partnership
    -0.62
     MSM
    -0.60
    SF
    -0.59
    mite
    -0.59
    POSITIVE LOGITS
     overseeing
    1.23
     protecting
    1.12
     safegu
    1.10
     preserving
    1.05
     ensuring
    1.04
     maintaining
    1.04
     upholding
    1.02
     navigating
    1.01
     keeping
    1.00
     regulating
    1.00
    Activations Density 0.018%

    No Known Activations