© 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. 95462
    Prev
    Next
    INDEX
    Explanations

    multi-language text fragments

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

    This neuron detects news‐wire byline/dateline markers—abbreviations like “(AP)”, “St.”, “T.”, or numeric scores that typically appear in news agency headers.

    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
    ADING
    -1.05
    UBLE
    -0.96
    pozy
    -0.92
     kwi
    -0.91
     flèche
    -0.90
    iczna
    -0.89
    UMBUS
    -0.89
     couronne
    -0.88
    RESH
    -0.88
     campe
    -0.87
    POSITIVE LOGITS
     memutus
    1.20
     memperbaiki
    1.03
     включая
    0.94
     joaat
    0.92
     quartz
    0.91
     zarówno
    0.89
    </b>
    0.89
    ające
    0.88
     digestion
    0.88
     oorspron
    0.87
    Activations Density 0.008%

    No Known Activations