© 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. 14999
    Prev
    Next
    INDEX
    Explanations

    Q & A, code documentation

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

    The neuron detects JSDoc-style type annotation tokens (e.g. the “@param/@returns” tags, curly-brace type declarations, and “{*}”, “{Object}”, etc.).

    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
     their
    -1.89
     it
    -1.80
     people
    -1.69
     both
    -1.60
     all
    -1.55
     many
    -1.50
     only
    -1.47
     even
    -1.46
     from
    -1.45
     at
    -1.41
    POSITIVE LOGITS
    Ingredienti
    1.48
    튿
    1.48
     figliu
    1.45
     küche
    1.41
     schuhe
    1.38
     REALLY
    1.38
    Aantal
    1.34
    küche
    1.34
    really
    1.31
     เมื่อ
    1.31
    Activations Density 0.005%

    No Known Activations