© 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. 10-GEMMASCOPE-RES-131K
    4. 101834
    Prev
    Next
    INDEX
    Explanations

    instructions and actions

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

    Japanese verb stems and conjugations.

    oai_token-act-pair · claude-4-5-sonnetTriggered by @bfgcn7frf5

    The neuron responds to Japanese subword segments—especially contentful morphemes (verbs, nouns, etc.) in Japanese text.

    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
     not
    -1.59
     Вене
    -1.47
     by
    -1.45
     /
    -1.42
     \
    -1.40
     +
    -1.37
    ability
    -1.36
    ⇘
    -1.36
     >
    -1.35
     on
    -1.34
    POSITIVE LOGITS
     dificuldades
    1.81
     játék
    1.59
    
    1.52
     avulla
    1.47
     suppos
    1.45
    鋅
    1.45
    ขอบคุณ
    1.45
     vendeurs
    1.44
    我们需要
    1.43
    esercito
    1.42
    Activations Density 0.007%

    No Known Activations