© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 5247
    Prev
    Next
    INDEX
    Explanations

    single quotes or punctuation

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

    The neuron responds strongly to apostrophe characters (the single-quote or curly-quote token), i.e. the punctuation used in contractions, possessives, and similar quote-like uses.

    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_31_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
     reten
    0.71
    },}
    0.68
     connotations
    0.66
     slags
    0.65
     biofilms
    0.65
     permutations
    0.64
     rechercher
    0.64
    ту
    0.63
     ischa
    0.63
     plaintext
    0.63
    POSITIVE LOGITS
    s
    1.05
    𝐬
    0.88
    𝘀
    0.85
    nın
    0.84
    I
    0.83
    R
    0.82
    الله
    0.77
    G
    0.77
    𝗱
    0.77
    entrada
    0.77
    Activations Density 1.498%

    No Known Activations