© 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
    2. Gemma-2-27B
    3. 10-GEMMASCOPE-RES-131K
    4. 61166
    Prev
    Next
    INDEX
    Explanations

    how quickly, accurately, often, well

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

    The neuron activates on adverbs (especially manner or frequency adverbs like “quickly,” “efficiently,” “accurately,” “often,” “well”) to detect words describing how or how often something is done.

    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
    )$.
    -1.88
    );
    -1.74
     PUSTAKA
    -1.55
     amusing
    -1.52
     χρήση
    -1.50
    mathrm
    -1.47
    ir
    -1.45
     zahlreichen
    -1.45
    }
    -1.41
    是有
    -1.40
    POSITIVE LOGITS
     we
    2.58
     they
    2.19
     you
    2.05
     cewek
    1.73
     déput
    1.61
    ໊
    1.61
    事で
    1.59
    ORETICAL
    1.59
     it
    1.52
    ering
    1.52
    Activations Density 0.007%

    No Known Activations