© 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. 65000
    Prev
    Next
    INDEX
    Explanations

    up to, check back

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

    The neuron fires on the word “up” when it appears as a verbal particle in phrasal constructions (e.g. sell up, pay up, check up, change up).

    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
     aparato
    -2.03
     What
    -2.02
     Because
    -1.98
     Some
    -1.84
     Shortly
    -1.84
     from
    -1.83
     facilitate
    -1.80
     There
    -1.80
     fulfill
    -1.79
     September
    -1.73
    POSITIVE LOGITS
    
    2.05
     mauva
    2.02
     elegant
    1.88
    }}$
    1.88
    on
    1.81
    图源
    1.80
     astounding
    1.79
    extremely
    1.78
     momentous
    1.77
     kinda
    1.77
    Activations Density 0.033%

    No Known Activations