© 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. 70102
    Prev
    Next
    INDEX
    Explanations

    entities involved in competitive scenarios, such as opponents and actors in a narrative context

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

    people in interaction roles

    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-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
    :✨
    -0.72
    OGND
    -0.66
    webElementGuid
    -0.59
    RTDA
    -0.59
     oprot
    -0.57
    ፩
    -0.57
    Autoritní
    -0.57
    AnimationsModule
    -0.55
     húmedo
    -0.53
     estaciones
    -0.52
    POSITIVE LOGITS
     contr
    0.33
    autorelease
    0.33
    Personendaten
    0.32
     Welle
    0.31
     IDisposable
    0.30
     alma
    0.30
    qquad
    0.29
     arter
    0.28
     aff
    0.27
     schli
    0.27
    Activations Density 0.066%

    No Known Activations