© 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-3-12B
    5. 24-GEMMASCOPE-2-RES-16K
    7. 5055
    Prev
    Next
    INDEX
    Explanations

    word parts in multiple languages

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

    The neuron strongly responds to discourse‐level signal words—brief transition words (e.g. “then”), demonstrative determiners (e.g. “those”), and evaluative adjectives (e.g. “flattering”).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-pt/resid_post/layer_24_width_16k_l0_medium
    Prompts (Dashboard)
    392,802 prompts, 256 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    1
    0.84
    7
    0.82
    0
    0.80
     '
    0.77
    9
    0.72
    5
    0.71
    6
    0.69
    8
    0.68
    Rank
    0.68
    ['
    0.67
    POSITIVE LOGITS
     astăzi
    0.76
     dbcontext
    0.74
     voluptate
    0.73
     identificado
    0.71
     préférences
    0.70
    샹
    0.70
     artisanal
    0.70
     espero
    0.70
     genocide
    0.70
     görev
    0.70
    Activations Density 0.000%

    No Known Activations