© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 111813
    Prev
    Next
    INDEX
    Explanations

    FIN, IB, ML, PE, AD, KD

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

    The neuron fires on multi‐letter uppercase strings—that is, acronyms or initialisms.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
     avanzada
    -1.50
     misschien
    -1.38
     maybe
    -1.38
     talentos
    -1.29
     this
    -1.29
     when
    -1.29
     suelen
    -1.29
     each
    -1.27
     our
    -1.26
     belki
    -1.25
    POSITIVE LOGITS
    
    1.73
     pét
    1.50
     bougies
    1.44
    !“
    1.41
     liné
    1.40
    deauna
    1.38
     acide
    1.36
     cabine
    1.36
     saine
    1.36
    慙
    1.34
    Activations Density 0.157%

    No Known Activations