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. 26302
    Prev
    Next
    INDEX
    Explanations

    occurrences of the token "Tr" or variations of it, indicating a focus on a specific subject or identifier

    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
    128
    +0.11
    0.5%
    232
    +0.11
    0.5%
    706
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    128
    +0.11
    0.02
    232
    +0.11
    0.02
    213
    +0.11
    0.02
    Negative Logits
     yeast
    -0.71
    lihood
    -0.70
     arts
    -0.70
    WARE
    -0.69
    ELS
    -0.68
    ãģ®éŃĶ
    -0.67
    âĸ¬âĸ¬
    -0.67
     tong
    -0.66
     writ
    -0.66
     actionGroup
    -0.65
    POSITIVE LOGITS
    ainer
    1.29
    acer
    1.27
    ained
    1.27
    ains
    1.21
    istan
    1.21
    unks
    1.19
    ulia
    1.18
    umps
    1.17
    icol
    1.17
    ample
    1.14
    Activations Density 0.010%

    No Known Activations