© 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. 22-GEMMASCOPE-RES-131K
    4. 36495
    Prev
    Next
    INDEX
    Explanations

    lack of capability or inclusion

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

    This neuron highlights negations and words expressing limitations or drawbacks (e.g. “not,” “cannot,” “does not,” “unfortunately,” “disadvantage”).

    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
     cyclisme
    -1.13
     навіть
    -1.08
    isissez
    -1.06
    Ранее
    -1.02
     huéspedes
    -0.97
     connaissent
    -0.97
    PROTO
    -0.94
     libéral
    -0.92
     السياس
    -0.90
    ícias
    -0.90
    POSITIVE LOGITS
     support
    1.28
     protection
    1.20
     help
    1.13
     guarantee
    1.09
     flexibility
    1.09
     efficiency
    1.03
     efficient
    1.00
     access
    1.00
     information
    1.00
     protect
    0.98
    Activations Density 0.057%

    No Known Activations