© 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. 67612
    Prev
    Next
    INDEX
    Explanations

    introduces content below

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

    The neuron detects “Below is a … of …”–style lead-in phrases that introduce lists, summaries, copies, or other content blocks.

    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
     it
    -1.70
     this
    -1.34
     at
    -1.27
     ור
    -1.20
    ِن
    -1.19
    nehå
    -1.17
    romatic
    -1.15
    ရ
    -1.13
     effetti
    -1.12
     zamanda
    -1.11
    POSITIVE LOGITS
     list
    1.43
     VERY
    1.32
     فقط
    1.32
     OGSÅ
    1.31
     somewhat
    1.31
     roughly
    1.31
    タイプの
    1.30
     a
    1.27
     alguns
    1.25
     Edelstahl
    1.23
    Activations Density 0.034%

    No Known Activations