© 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. 35912
    Prev
    Next
    INDEX
    Explanations

    coding error and technical terms

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

    The neuron activates on programming‐language identifiers (e.g. variable names like “self” and class/member names like “Vertical”), i.e. tokens that are code identifiers.

    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
     to
    -1.27
     only
    -1.19
    用に
    -1.11
     result
    -1.05
     come
    -1.04
     капю
    -1.03
    Who
    -1.00
    phans
    -1.00
     فقط
    -0.99
    aleza
    -0.98
    POSITIVE LOGITS
     вклад
    1.12
    鐽
    1.00
     dormitorios
    0.96
     trám
    0.95
     logements
    0.94
     láp
    0.94
    我們
    0.94
    ⽉
    0.93
    衣裙
    0.93
    我们
    0.93
    Activations Density 0.000%

    No Known Activations