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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & 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. 1567
    Prev
    Next
    INDEX
    Explanations

    references to formal documents and statements regarding governance, financial transactions, and personal experiences

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

    Followed by punctuation or a stop word

    np_acts-logits-general · gemini-2.0-flash

    common words followed by specific context

    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
     esprits
    -0.36
    else
    -0.34
    the
    -0.33
     señales
    -0.32
     palabras
    -0.32
    jącej
    -0.32
     Octubre
    -0.31
     advances
    -0.31
    laştır
    -0.31
     počas
    -0.31
    POSITIVE LOGITS
     &___
    0.75
    aarrggbb
    0.71
    CppMethod
    0.71
    URLException
    0.71
     propOrder
    0.70
    BagLayout
    0.68
     Normdatei
    0.68
    @[+][
    0.68
    principalColumn
    0.66
     surla
    0.66
    Activations Density 0.589%

    No Known Activations