© 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
    3. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    5. Gemma-2-9B-IT
    7. Residual Stream - 131k
    9. 20-GEMMASCOPE-RES-131K
    11. 82328
    Prev
    Next
    INDEX
    Explanations

    verbs and related adjuncts indicating capability and potential action

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

    common word + connector

    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
     Klage
    -0.33
     AssemblyVersion
    -0.32
     Miß
    -0.32
    tryk
    -0.31
     Erfahrungen
    -0.30
    verband
    -0.30
     myths
    -0.30
    렸
    -0.30
    MeasureSpec
    -0.30
     चाहिए
    -0.29
    POSITIVE LOGITS
    OGND
    0.86
    ########.
    0.85
    BeginContext
    0.74
     Taktlose
    0.67
    TagMode
    0.67
    setVerticalGroup
    0.67
     propOrder
    0.65
    :✨
    0.63
     surla
    0.63
    rungsseite
    0.62
    Activations Density 0.453%

    No Known Activations