© 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. 28252
    Prev
    Next
    INDEX
    Explanations

    may, could, will, some, any

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

    The neuron is keyed to conditional or hedging language—words and short multi-word cues like “except,” “while,” “may/might,” “need to,” “could,” etc., that introduce caveats, qualifications, or trial/suggestion contexts.

    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
     this
    -1.18
     descendre
    -1.16
     suche
    -1.15
     până
    -1.13
     poursuivre
    -1.13
     horri
    -1.09
     aceste
    -1.09
    不仅
    -1.09
     также
    -1.06
     these
    -1.05
    POSITIVE LOGITS
     slightly
    2.09
     немного
    1.84
     slight
    1.81
     occasionally
    1.77
     некоторых
    1.70
     some
    1.70
     occasional
    1.68
    Some
    1.63
     algunas
    1.63
     allerdings
    1.60
    Activations Density 0.039%

    No Known Activations