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. 12122
    Prev
    Next
    INDEX
    Explanations

    URLs or web links

    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
    373
    +0.10
    0.5%
    445
    +0.10
    0.5%
    385
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    445
    +0.10
    0.02
    439
    +0.10
    0.02
    449
    +0.10
    0.02
    Negative Logits
     overshadow
    -0.73
     mainland
    -0.63
     succeeding
    -0.63
     rooting
    -0.62
     Labrador
    -0.62
     retali
    -0.61
     extinction
    -0.60
     performing
    -0.60
     sway
    -0.59
     proportion
    -0.59
    POSITIVE LOGITS
    ://
    1.91
    www
    1.29
    :/
    1.11
     www
    1.04
    natureconservancy
    1.04
    youtu
    1.02
    docs
    0.96
    :\
    0.91
    ww
    0.90
    tiny
    0.86
    Activations Density 0.008%

    No Known Activations