© 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. 1731
    Prev
    Next
    INDEX
    Explanations

    holding back

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

    The neuron flags verbs and verb forms that express withholding or restraining (e.g., hold, hold back, withhold, holding).

    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
     ingrained
    -2.06
     undeniably
    -1.84
     dazzling
    -1.84
     heightened
    -1.82
     puzzling
    -1.80
     time
    -1.77
     remarkable
    -1.76
     ребенку
    -1.76
     Surprisingly
    -1.74
     that
    -1.69
    POSITIVE LOGITS
    gged
    2.05
    も含
    1.97
    ArgsConstructor
    1.91
    ).
    1.88
    s
    1.84
    狴
    1.82
    8
    1.77
    ériques
    1.74
    cisive
    1.73
    ౮
    1.73
    Activations Density 0.003%

    No Known Activations