© 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. 34-GEMMASCOPE-RES-131K
    7. 769
    Prev
    Next
    INDEX
    Explanations

    can be

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

    The neuron detects the modal verb “can” (especially in “can be” constructions) signaling possibility or capability.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
     inay
    -1.03
     will
    -0.98
    locene
    -0.95
     ポーズ
    -0.93
     xxiii
    -0.93
    らなかった
    -0.92
    یی
    -0.89
     confider
    -0.89
    BIS
    -0.88
     XXXV
    -0.88
    POSITIVE LOGITS
     be
    3.11
     sometimes
    2.75
    Sometimes
    2.16
     often
    2.14
     parfois
    1.96
     Sometimes
    1.91
     manchmal
    1.86
    sometimes
    1.83
     иногда
    1.80
     være
    1.72
    Activations Density 0.068%

    No Known Activations