© 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. 71428
    Prev
    Next
    INDEX
    Explanations

    to followed by verb

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

    The neuron strongly activates on the infinitive marker “to” (especially in verb-phrase sequences like “trying to,” “tried to,” “learned to,” etc.).

    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
     fatale
    -1.28
     ($('#
    -1.21
     vitale
    -1.16
     brune
    -1.14
     tirs
    -1.13
     herbes
    -1.12
     regionale
    -1.09
    lieving
    -1.09
     alho
    -1.09
     bestellen
    -1.05
    POSITIVE LOGITS
    /***
    
    1.41
     understand
    1.27
    lový
    1.22
    GMENT
    1.16
     ktoré
    1.16
     květ
    1.16
     transverse
    1.13
     Perubahan
    1.13
     które
    1.13
     accept
    1.13
    Activations Density 0.016%

    No Known Activations