© 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 - 16k
    5. 20-GEMMASCOPE-RES-16K
    6. 4405
    Prev
    Next
    INDEX
    Explanations

    modal verbs indicating possibility or necessity

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

    Follows "could" or "would" or "should"

    np_acts-logits-general · gemini-2.0-flash

    Modal perfect past constructions

    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-16k
    Configuration
    google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    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
    guiente
    -0.54
     vôtre
    -0.42
     chrétien
    -0.42
     nôtre
    -0.40
     StatelessWidget
    -0.39
     règne
    -0.38
    arrête
    -0.38
    cías
    -0.38
     Kini
    -0.37
     vœ
    -0.37
    POSITIVE LOGITS
     have
    0.97
     had
    0.93
    have
    0.91
    Have
    0.82
     having
    0.82
     been
    0.82
     Have
    0.79
     Hassel
    0.73
    had
    0.73
     avuto
    0.71
    Activations Density 0.146%

    No Known Activations