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

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 35128
    Prev
    Next
    INDEX
    Explanations

    verbs and prepositions

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

    The neuron tends to light up on tokens that serve as discourse/connective markers or key content words—especially verbs and complementizers that introduce or connect clauses (e.g. “on,” “that,” “accepted,” “read,” “proved,” “agreed”).

    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
    ξύ
    -1.06
     certain
    -1.03
    キニ
    -0.99
    什么样的
    -0.98
     totul
    -0.97
    tainty
    -0.95
     filtros
    -0.94
    pisode
    -0.93
     who
    -0.93
    üglich
    -0.92
    POSITIVE LOGITS
    ‏‏
    0.97
     this
    0.96
     Duca
    0.95
    }
    0.90
     prawda
    0.88
     the
    0.88
    ytä
    0.86
    了出来
    0.85
    คำ
    0.84
    croft
    0.84
    Activations Density 0.018%

    No Known Activations