© 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. 97167
    Prev
    Next
    INDEX
    Explanations

    specific technical terms and phrases related to programming or coding languages

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

    mathematical expressions and methods

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

    It detects tokens that are part of structured metadata/technical headings or codes—all-caps labels, acronyms, numeric codes and similar table-like or header tokens.

    oai_token-act-pair · gpt-5-miniTriggered 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
     CURIAM
    -0.50
    Bakgrunnsstoff
    -0.47
    oplayer
    -0.45
    rrggbb
    -0.42
    OrBuilder
    -0.41
    +#+#
    -0.39
     picioare
    -0.39
     prochaines
    -0.38
    IntoConstraints
    -0.37
    dega
    -0.37
    POSITIVE LOGITS
    LookAnd
    0.65
     للاسماء
    0.59
    httphttps
    0.55
     Normdatei
    0.52
     ***!
    0.50
     cherchés
    0.49
     jadx
    0.47
    ⓧ
    0.47
    tagHelperRunner
    0.47
     Réponses
    0.46
    Activations Density 15.043%

    No Known Activations