© 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. 10-GEMMASCOPE-RES-131K
    7. 21801
    Prev
    Next
    INDEX
    Explanations

    encoding errors and special characters

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

    The neuron flags occurrences of mis-decoded or garbled non-ASCII characters (e.g. the recurring “â” artifacts).

    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
     Surprisingly
    -3.17
     unrelenting
    -3.08
     Quite
    -2.98
     perplexing
    -2.92
     burgeoning
    -2.86
     bolster
    -2.84
     Since
    -2.84
     enhances
    -2.81
     While
    -2.80
     Taking
    -2.80
    POSITIVE LOGITS
     Dinas
    3.14
    
    3.13
     delightfully
    3.05
     unbelievably
    3.00
     insanely
    3.00
     ridiculously
    2.91
     zoude
    2.91
     amazingly
    2.84
    Ꭳ
    2.80
     wonderfully
    2.77
    Activations Density 0.002%

    No Known Activations