© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 79622
    Prev
    Next
    INDEX
    Explanations

    references to specific entities and relationships, particularly involving legal or formal language

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

    Apache way

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

    the neuron detects tokens that are proper names, all-caps identifiers, acronyms, or numeric/technical tokens typical of scientific/technical text.

    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
     comuniques
    -0.47
    EndContext
    -0.39
    IVEREF
    -0.39
    ⓧ
    -0.39
    GEBURTSDATUM
    -0.38
    transQ
    -0.38
    Попис
    -0.37
    iNdEx
    -0.36
     Unab
    -0.36
    prehension
    -0.35
    POSITIVE LOGITS
    XmlAccessType
    0.54
     ErrIntOverflow
    0.50
     bacio
    0.47
     Lux
    0.46
     kasarigan
    0.46
    ViewImports
    0.45
    webElementGuid
    0.44
     bribes
    0.44
    tagHelper
    0.43
     Touch
    0.43
    Activations Density 0.760%

    No Known Activations