© 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. 40172
    Prev
    Next
    INDEX
    Explanations

    test selectors and imports

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

    The neuron is picking up on structural code delimiters and punctuation (e.g. underscores, colons, dots, equals signs, @-prefixes) rather than natural-language content.

    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
    が可能
    -1.21
     绘画
    -1.20
    aceite
    -1.16
    ASTM
    -1.13
     männ
    -1.12
    attente
    -1.09
    pitre
    -1.08
    さっき
    -1.08
    相変わらず
    -1.08
     новом
    -1.05
    POSITIVE LOGITS
     también
    1.17
    чи
    1.16
     такое
    1.14
     к
    1.12
     إلى
    1.10
     с
    1.09
    ме
    1.08
     TODO
    1.07
     że
    1.07
    /**
    
    1.05
    Activations Density 0.210%

    No Known Activations