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. 26273
    Prev
    Next
    INDEX
    Explanations

    repeated instances of the article "the"

    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
    393
    +0.19
    0.9%
    481
    +0.18
    0.9%
    636
    +0.13
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    362
    +0.19
    0.08
    636
    +0.18
    0.06
    393
    +0.13
    0.06
    Negative Logits
    osponsors
    -0.78
    handedly
    -0.69
    govtrack
    -0.63
    Turkey
    -0.63
     accompl
    -0.63
    insk
    -0.62
    pers
    -0.61
    @@
    -0.60
    ato
    -0.58
    ofi
    -0.58
    POSITIVE LOGITS
     same
    1.45
    ses
    1.20
     dreaded
    1.09
     latter
    1.04
     aforementioned
    1.04
     slightest
    1.01
    same
    1.00
     latest
    1.00
     brunt
    0.99
     highest
    0.94
    Activations Density 0.253%

    No Known Activations