© 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
    2. Gemma-2-27B
    3. 10-GEMMASCOPE-RES-131K
    4. 61088
    Prev
    Next
    INDEX
    Explanations

    might / will / have

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

    The neuron detects hedging or generalizing qualifiers—words like “some,” “many,” or “might argue” that introduce opinions or broad generalizations.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
    From
    -1.84
     &
    -1.77
    Our
    -1.73
    Your
    -1.71
    他にも
    -1.66
     cómod
    -1.64
     Deletes
    -1.62
    What
    -1.59
     any
    -1.58
    さんも
    -1.57
    POSITIVE LOGITS
    ༘
    1.85
     grandeza
    1.73
    丷
    1.71
    ⢅
    1.68
     înal
    1.66
    стойчи
    1.66
    🡺
    1.63
    珦
    1.63
    ⩤
    1.63
     themselves
    1.61
    Activations Density 0.017%

    No Known Activations