© 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. 59793
    Prev
    Next
    INDEX
    Explanations

    may/might expressing possibility

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

    The neuron activates most strongly on modal verbs expressing possibility or uncertainty (e.g. “may,” “might,” “could”).

    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
     he
    -1.33
    レーター
    -1.32
     値下げ
    -1.25
     JAKARTA
    -1.16
     all
    -1.16
     toks
    -1.14
    w
    -1.13
     tuturor
    -1.12
    enaire
    -1.12
    urb
    -1.12
    POSITIVE LOGITS
     Это
    1.53
     там
    1.34
    大切な
    1.27
     інших
    1.26
     diisi
    1.25
     Käufer
    1.24
     agradec
    1.21
    中略
    1.21
     contrato
    1.20
     postępo
    1.20
    Activations Density 0.076%

    No Known Activations