© 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. 10-GEMMASCOPE-RES-131K
    7. 72749
    Prev
    Next
    INDEX
    Explanations

    arguing or contending

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

    This neuron specifically detects occurrences of the verb “contend” (including its “contends” form), flagging when someone is making an argument or assertion.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
     worin
    -2.86
    }
    -2.48
    省市镇
    -2.45
    鲠
    -2.42
    gingen
    -2.34
    鉨
    -2.34
     Gründung
    -2.31
    ጺ
    -2.27
    vantaged
    -2.25
    鷂
    -2.23
    POSITIVE LOGITS
    ↵
    3.11
    </em>
    3.03
    橚
    2.86
     самый
    2.75
     blazing
    2.58
     sorta
    2.47
     they
    2.36
     staggering
    2.36
     craz
    2.34
    蓠
    2.34
    Activations Density 0.003%

    No Known Activations