Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-9B-IT
    4. Residual Stream - 16k
    5. 20-GEMMASCOPE-RES-16K
    6. 1026
    Prev
    Next
    INDEX
    Explanations

    references to mathematical operations and scientific notations

    oai_token-act-pair · gpt-4o-miniTriggered by @bot

    Mathematical simplification or parallel computation

    np_acts-logits-general · gemini-2.0-flash

    simplify mathematical expressions or parallel processing tasks

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.20.hook_resid_post
    Hook Layer
    20
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    TagMode
    -0.83
    таратура
    -0.71
    principalColumn
    -0.71
    tagHelperRunner
    -0.71
     CreateTagHelper
    -0.71
    RegressionTest
    -0.70
     cherchés
    -0.70
     BoxFit
    -0.69
    postsleuth
    -0.67
    Архівовано
    -0.66
    POSITIVE LOGITS
     Ortsteil
    0.38
     llama
    0.34
     mengin
    0.32
    DataItem
    0.32
     Mexique
    0.31
     men
    0.31
     dokon
    0.31
     réve
    0.31
    Men
    0.31
     Famille
    0.31
    Activations Density 0.186%

    No Known Activations