© 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-IT
    4. Residual Stream - 131k
    5. 20-GEMMASCOPE-RES-131K
    6. 126137
    Prev
    Next
    INDEX
    Explanations

    items and actions related to food and entertainment

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

    language or specific names

    np_acts-logits-general · gemini-2.5-flash-lite

    sensational, hyperbolic, or absurd phrasing that uses vivid, attention-grabbing references (often humorous or shocking).

    oai_token-act-pair · gpt-5Triggered by @vetterc0
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k
    Configuration
    google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    131,072
    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
     char
    -0.30
    Descripció
    -0.29
     militer
    -0.28
    chargeable
    -0.28
     dele
    -0.28
    gebene
    -0.27
     “
    -0.26
     pooh
    -0.26
     miłości
    -0.26
     pasi
    -0.26
    POSITIVE LOGITS
    EDEFAULT
    0.74
    دانشنامهٔ
    0.66
     ویکی‌پدی
    0.65
    featureID
    0.65
    évaluateur
    0.61
     Infórmanos
    0.60
     pinulongan
    0.58
     utafitiHapana
    0.57
     invokingState
    0.56
    BagLayout
    0.55
    Activations Density 0.045%

    No Known Activations