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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-9B
    4. Residual Stream - 16k
    5. 28-GEMMASCOPE-RES-16K
    6. 3924
    Prev
    Next
    INDEX
    Explanations

    programming-related function definitions and operations

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

    Code snippets containing "return" or variables like "m_" or "p"

    np_acts-logits-general · gemini-2.0-flash

    PHP, Javascript, MATLAB code keywords

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B @ 28-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-pt-res/layer_28/width_16k/average_l0_119
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.28.hook_resid_post
    Hook Layer
    28
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ⓧ
    -0.42
    horabuena
    -0.34
     actuelles
    -0.32
     ahor
    -0.32
     Actualmente
    -0.31
     ahorro
    -0.31
     totality
    -0.31
     éditoriale
    -0.30
     confiance
    -0.29
    éndole
    -0.29
    POSITIVE LOGITS
    culate
    0.60
     müſſen
    0.58
     zuſammen
    0.58
     Geſch
    0.57
     geſch
    0.55
     laſſen
    0.54
     undergo
    0.52
    ftn
    0.52
     ſelbſt
    0.52
    await
    0.52
    Activations Density 0.761%

    No Known Activations