© 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. 70931
    Prev
    Next
    INDEX
    Explanations

    code keywords and separators

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

    The neuron fires on log‐message tokens—especially log‐level labels (DEBUG, INFO, WARNING) and other uppercase words in logging calls.

    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
     Heilige
    -1.47
     jonge
    -1.38
     cordón
    -1.34
     heilige
    -1.30
     you
    -1.28
     mężczy
    -1.22
     weißen
    -1.21
    łeś
    -1.21
    IPV
    -1.20
     the
    -1.20
    POSITIVE LOGITS
     with
    1.95
     from
    1.74
     without
    1.44
     at
    1.43
    ancang
    1.33
     despite
    1.31
     avoiding
    1.31
     after
    1.28
     amidst
    1.24
     before
    1.20
    Activations Density 0.008%

    No Known Activations