© 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. 25352
    Prev
    Next
    INDEX
    Explanations

    letter followed by hyphen or shape

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

    This neuron detects standalone uppercase single-letter tokens (e.g. “K”, “B”, “I”, “G”) used as labels or identifiers.

    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
     
    -1.63
     chefe
    -1.57
     menino
    -1.55
     afirmar
    -1.52
     senhor
    -1.48
     the
    -1.45
     diretor
    -1.43
    ၟ
    -1.42
     advogado
    -1.41
     irmão
    -1.41
    POSITIVE LOGITS
     carteles
    1.59
    erializer
    1.49
     tradis
    1.48
     magasiner
    1.47
     tamaños
    1.46
     寸
    1.45
     videojuegos
    1.41
     peluche
    1.38
     deportes
    1.38
    "../
    1.35
    Activations Density 0.102%

    No Known Activations