© 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. 123502
    Prev
    Next
    INDEX
    Explanations

    hone, home, sharpen

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

    The neuron strongly activates on occurrences of the letter sequence "hone" (e.g. hone, honed, honing), regardless of context.

    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
    </h3>
    -2.84
    谖
    -2.28
     you
    -2.09
     которы
    -2.08
     Auffassung
    -2.05
     Who
    -2.02
    嗵
    -1.99
    _{\
    -1.99
    髏
    -1.99
    )<\
    -1.97
    POSITIVE LOGITS
     varieties
    2.38
     但
    2.30
    with
    2.25
     akin
    2.08
     但是
    2.03
     είναι
    2.03
    that
    2.02
    沨
    2.02
    里
    2.02
    +.
    1.98
    Activations Density 0.002%

    No Known Activations