© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 65594
    Prev
    Next
    INDEX
    Explanations

    technical descriptions

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

    The neuron fires on unusually long tokens (e.g. identifiers or comment words) in code, effectively detecting long multi-character words.

    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
    -1.21
     Feuchtigkeit
    -1.21
     allerlei
    -1.16
     heilige
    -1.14
    your
    -1.08
     harten
    -1.06
     spannend
    -1.06
    ____
    -1.05
    either
    -1.05
    ___________
    -1.05
    POSITIVE LOGITS
     despite
    1.48
     will
    1.42
     since
    1.35
     only
    1.32
     regardless
    1.28
     just
    1.23
     even
    1.20
     although
    1.18
     if
    1.15
     without
    1.14
    Activations Density 0.012%

    No Known Activations