© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Gemma-2-27B
    3. 22-GEMMASCOPE-RES-131K
    4. 38930
    Prev
    Next
    INDEX
    Explanations

    how to ask questions

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

    The neuron fires strongly on title or heading tokens that form question or instructional phrases (e.g. “How to…,” “What to…,” “Why…”) in headings or article titles.

    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
     this
    -1.54
     soldadura
    -1.27
     Posteriormente
    -1.26
    ἐ
    -1.17
    
    -1.16
     Erfüllung
    -1.14
     Veh
    -1.14
     maleta
    -1.14
     aktivitet
    -1.14
    unción
    -1.13
    POSITIVE LOGITS
     and
    1.94
     to
    1.91
    </h2>
    1.66
     you
    1.41
     می‌توانید
    1.33
    </h1>
    1.30
     for
    1.26
    </strong>
    1.23
    4
    1.23
    </h3>
    1.22
    Activations Density 0.004%

    No Known Activations