© 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. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 104361
    Prev
    Next
    INDEX
    Explanations

    grammatical forms and sentence types

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

    The neuron detects explicit mentions or discussions of “past tense” constructions—e.g. the phrase “in + past tense of a verb.”

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    试图
    -1.20
     cliquer
    -1.16
    攢
    -1.13
    ㈯
    -1.12
     torino
    -1.10
    kissen
    -1.09
    進去
    -1.08
     sanciones
    -1.08
     rosario
    -1.07
    弊社
    -1.07
    POSITIVE LOGITS
     of
    1.75
    mpang
    1.18
     in
    1.13
    Beskrivelse
    1.05
     used
    1.04
    ↵
    1.04
     verbs
    1.03
    nextProps
    1.02
    asci
    1.02
     toujours
    1.00
    Activations Density 0.019%

    No Known Activations