© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 25160
    Prev
    Next
    INDEX
    Explanations

    Middle earth, Harry Potter, Korean names

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

    The neuron fires on capitalized proper names and titles—especially character names and story headings.

    oai_token-act-pair · o4-mini
    Triggered 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
    Negative Logits
     this
    -1.45
     last
    -1.22
    adeira
    -1.17
     amig
    -1.13
     Testamento
    -1.13
     previous
    -1.12
     边框
    -1.03
    ampiyon
    -1.03
     Amigos
    -1.03
     rosa
    -1.02
    POSITIVE LOGITS
     trám
    1.36
    因為
    1.28
    Är
    1.21
     которы
    1.14
     Fähig
    1.13
     rám
    1.13
     passende
    1.13
    Questão
    1.13
     perfekte
    1.12
     }))
    1.11
    Activations Density 0.014%

    No Known Activations