© 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. 22-GEMMASCOPE-RES-131K
    4. 95366
    Prev
    Next
    INDEX
    Explanations

    acronyms and initialisms

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

    The neuron fires on tokens that are uppercase letter sequences immediately followed by digits (e.g. “SGE532”, “WML516”, “KKY540”), i.e. three‐letter acronyms with numeric suffixes.

    oai_token-act-pair · o4-miniTriggered 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
     a
    -2.06
     offenbar
    -1.67
    為に
    -1.58
     aclarar
    -1.58
     zweifel
    -1.54
     erwähnt
    -1.54
     zufried
    -1.48
    .
    -1.48
    parrot
    -1.47
     the
    -1.46
    POSITIVE LOGITS
    s
    1.96
    かもしれませんが
    1.68
     kutu
    1.59
     boya
    1.52
     سایر
    1.49
     maksi
    1.48
     katk
    1.48
     puedo
    1.45
    ですが
    1.45
     tekno
    1.45
    Activations Density 0.183%

    No Known Activations