© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-9B
    4. Residual Stream - 16k
    5. 11-GEMMASCOPE-RES-16K
    6. 4994
    Prev
    Next
    INDEX
    Explanations

    instances of the word "hold" and related concepts indicating possession or responsibility

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B @ 11-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-pt-res/layer_11/width_16k/average_l0_118
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.11.hook_resid_post
    Hook Layer
    11
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    CppMethod
    -0.59
    Personensuche
    -0.56
     Ames
    -0.56
    astify
    -0.54
    الحياه
    -0.53
    ValueStyle
    -0.53
     CommonModule
    -0.52
     ***!
    -0.51
    expandindo
    -0.50
    simpleType
    -0.48
    POSITIVE LOGITS
     hold
    1.64
    Hold
    1.62
     Hold
    1.60
     Holding
    1.59
    hold
    1.58
    Holding
    1.58
     holds
    1.56
     held
    1.54
     Holds
    1.52
    holding
    1.51
    Activations Density 0.112%

    No Known Activations