© 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. 22-GEMMASCOPE-RES-131K
    7. 36270
    Prev
    Next
    INDEX
    Explanations

    Courtney Todd Photography

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

    The neuron fires on numeric tokens—digits, multi‐digit numbers (years, dates, phone numbers, counts, etc.).

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

    formatting elements and metadata in web page content, particularly dates, contact information, section headers, and structural markup.

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

    structured, non-sentential page elements such as numbers and dates, titles/headings and all-caps labels, punctuation/symbols, and contact/link strings typical of forms, ads, and site metadata.

    oai_token-act-pair · gpt-5Triggered 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
    ÈNE
    -1.33
     것이
    -1.17
    Figura
    -1.17
     něj
    -1.17
     لكن
    -1.16
    Meanwhile
    -1.14
     notamment
    -1.14
     trochu
    -1.11
    Fonte
    -1.10
    车站
    -1.10
    POSITIVE LOGITS
    ɗ
    1.29
     軽量
    1.18
     sanat
    1.18
     yrs
    1.17
     suaminya
    1.16
     tomen
    1.16
     обратить
    1.16
     сахара
    1.14
     折りたたみ
    1.13
    ¿½
    1.13
    Activations Density 0.070%

    No Known Activations