© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 16021
    Prev
    Next
    INDEX
    Explanations

    saying sorry

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

    The neuron is detecting apology phrases, notably various occurrences of “sorry.”

    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
    Ꮞ
    -2.55
    附近的
    -2.20
    ၟ
    -2.06
     kuris
    -1.89
    掼
    -1.88
    𖥦
    -1.85
    kurtka
    -1.82
    蒾
    -1.79
     które
    -1.77
    ၞ
    -1.77
    POSITIVE LOGITS
     autres
    2.14
    鲼
    2.11
    胸口
    1.79
     другие
    1.77
    齶
    1.76
     environ
    1.75
    us
    1.73
    toit
    1.73
    will
    1.67
    只剩下
    1.66
    Activations Density 0.003%

    No Known Activations