© 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. 130250
    Prev
    Next
    INDEX
    Explanations

    maleonitriledithiolate

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

    terms from non-Western languages, religious texts, or technical terminology with unusual letter combinations.

    oai_token-act-pair · claude-3-7-sonnet-20250219Triggered by @neilrathi

    The neuron lights up on low‐frequency, domain‐specific word fragments—especially parts of proper names, titles, or technical terms.

    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
     gaben
    -2.08
     ersche
    -2.02
     You
    -1.98
    kwür
    -1.91
    潆
    -1.87
    鸲
    -1.85
     wuß
    -1.82
     ваша
    -1.80
    onomian
    -1.80
    ଢ
    -1.74
    POSITIVE LOGITS
     about
    1.91
    []):
    1.84
    ↵↵↵↵↵↵↵↵↵↵
    1.69
     nedeniyle
    1.69
    ”,
    1.68
    ↵↵↵↵↵↵↵↵↵↵↵
    1.68
    u
    1.66
    ↵↵↵↵↵↵↵↵↵↵↵↵↵
    1.65
     futuros
    1.65
    m
    1.62
    Activations Density 0.002%

    No Known Activations