© 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-3-12B
    3. 24-GEMMASCOPE-2-RES-16K
    4. 12307
    Prev
    Next
    INDEX
    Explanations

    a bit nuanced

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

    The neuron is especially sensitive to common linking verbs—forms of “to be” (is, are) and related auxiliaries like “has.”

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-pt/resid_post/layer_24_width_16k_l0_medium
    Prompts (Dashboard)
    392,802 prompts, 256 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     pampered
    0.84
    হ্ন
    0.81
    と感じ
    0.79
    с
    0.79
     chances
    0.77
    ры
    0.76
    ",'
    0.76
     casado
    0.75
    лама
    0.74
    ছি
    0.73
    POSITIVE LOGITS
     Headers
    0.74
     Artwork
    0.74
    Lieutenant
    0.73
    MET
    0.73
    藝
    0.71
    optional
    0.71
    Long
    0.70
    Lect
    0.67
     پاسخ
    0.66
    og
    0.66
    Activations Density 0.001%

    No Known Activations