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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Gemma-2-27B
    3. 10-GEMMASCOPE-RES-131K
    4. 69337
    Prev
    Next
    INDEX
    Explanations

    believes in

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

    The neuron strongly activates on third-person singular present-tense English verbs (i.e. verbs ending in “-s” like “loves,” “believes,” “humiliates,” “gives,” 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_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
     tradicionais
    -1.73
     pessoais
    -1.63
     sociais
    -1.62
     elétrico
    -1.55
     maa
    -1.55
     DAN
    -1.49
    Kolor
    -1.48
    Ayrıca
    -1.48
    歺
    -1.46
     eier
    -1.40
    POSITIVE LOGITS
     knows
    1.57
     himself
    1.52
     still
    1.48
                
    1.43
     their
    1.34
     makes
    1.30
     ive
    1.30
    未満
    1.29
    στη
    1.29
     zato
    1.27
    Activations Density 0.009%

    No Known Activations