© 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. 71874
    Prev
    Next
    INDEX
    Explanations

    conjunctions and prepositions introducing clauses

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

    This neuron detects discourse‐linking words or phrases that introduce subordinate clauses—especially conditional or temporal connectors (e.g. as, if, once, when, after, pending, now).

    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
    même
    -1.19
    zzled
    -1.01
     aldrig
    -0.91
     système
    -0.89
    issez
    -0.88
    lucky
    -0.88
    ÇÕES
    -0.88
    Locker
    -0.87
    ariah
    -0.86
    combine
    -0.86
    POSITIVE LOGITS
     needed
    1.40
     better
    1.20
     needs
    1.16
     need
    1.11
     found
    1.09
     deemed
    1.08
     felt
    1.08
     demand
    1.04
     strength
    1.03
     get
    1.00
    Activations Density 0.057%

    No Known Activations