© 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. 79654
    Prev
    Next
    INDEX
    Explanations

    undef HAVE

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

    The neuron fires on tokens that are code identifiers or API names (especially uppercase or CamelCase symbols like CLR, Reset, Clone) and related punctuation constructs in code or code-like documentation.

    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
    琬
    -1.23
    
    -1.13
    ſta
    -1.12
    riy
    -1.11
    蛱
    -1.09
    länd
    -1.08
    ně
    -1.07
     ส์
    -1.05
     √
    -1.02
    География
    -1.01
    POSITIVE LOGITS
    or
    0.99
     poświę
    0.92
    and
    0.90
     altogether
    0.90
    comple
    0.88
    飾
    0.88
    古风
    0.86
    开来
    0.85
     removing
    0.85
    优秀的
    0.85
    Activations Density 0.000%

    No Known Activations