© 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. 28959
    Prev
    Next
    INDEX
    Explanations

    references to legal cases and specific legal terminology

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

    case report

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

    The neuron detects numeric tokens and numeric expressions (especially ages, years, and other numbers in clinical/legal/report contexts).

    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
     estekak
    -0.57
    ckså
    -0.53
     pylab
    -0.51
    zionalità
    -0.50
     fVar
    -0.50
    сылкі
    -0.49
    tionnaire
    -0.49
    AndroidJUnit
    -0.48
     nameof
    -0.48
     fontWeight
    -0.47
    POSITIVE LOGITS
    Portail
    0.47
     الحره
    0.41
    andExpect
    0.40
    tagHelperRunner
    0.39
    новништво
    0.38
     initComponents
    0.37
     trekken
    0.37
    InstanceState
    0.35
    principalColumn
    0.35
     propOrder
    0.35
    Activations Density 0.024%

    No Known Activations