© 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. 40141
    Prev
    Next
    INDEX
    Explanations

    how to answer questions

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

    The neuron detects question-style or instructional headings, i.e. tokens signaling “how to…,” “what is…,” “how much…,” “steps to…,” “methods to…,” “signs…,” etc.

    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
    ఝ
    -1.17
    unno
    -1.09
    Chaque
    -1.04
    年轻
    -1.02
    听到
    -0.99
     paciencia
    -0.98
    athered
    -0.98
    この
    -0.96
    时尚
    -0.96
     этих
    -0.95
    POSITIVE LOGITS
     which
    1.06
     and
    0.96
    ogues
    0.94
     to
    0.94
     '=',
    0.93
     in
    0.91
    顶部
    0.87
     hưng
    0.87
    టు
    0.86
     😘
    0.85
    Activations Density 0.056%

    No Known Activations