© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 2320
    Prev
    Next
    INDEX
    Explanations

    escalation

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

    The neuron fires strongly on tokens in non-English scripts or unusual foreign transliterations (e.g. Sinhala characters or romanized Hindi syllables).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    non-English text, particularly from languages using non-Latin scripts (such as Devanagari, Sinhala, Turkish, Catalan, Estonian, and other scripts).

    oai_token-act-pair · claude-4-5-sonnetTriggered by @jyhe0408

    non-English text, especially tokens from non-Latin scripts or Latin words with diacritics and other language-specific character patterns.

    oai_token-act-pair · gpt-5Triggered by @jyhe0408

    non-English or foreign-language text, especially tokens with non-ASCII characters, diacritics, or unusual transliterated consonant clusters.

    oai_token-act-pair · gpt-5Triggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-pt/resid_post/layer_31_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
     personable
    0.79
     dienst
    0.76
     valido
    0.73
     arguably
    0.72
     ratione
    0.71
     собствен
    0.70
     autént
    0.70
     weltweit
    0.69
     consigned
    0.69
     erziel
    0.68
    POSITIVE LOGITS
    ing
    0.89
    खोरी
    0.86
     کردن
    0.83
    áním
    0.82
    ING
    0.81
    거리
    0.78
     상태
    0.77
    ingar
    0.75
    ંટ
    0.75
    alk
    0.74
    Activations Density 0.085%

    No Known Activations