© 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
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 70645
    Prev
    Next
    INDEX
    Explanations

    dictates or allows

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

    The neuron strongly activates on relatively uncommon, content-rich words (technical terms, domain-specific nouns, or other low-frequency “informative” tokens) rather than on common function words.

    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
    dann
    -1.30
     signifikan
    -1.27
    isait
    -1.19
    stande
    -1.18
    krebs
    -1.17
    Banyak
    -1.14
    Bakgrunns
    -1.13
    Também
    -1.12
    Được
    -1.12
    vielen
    -1.11
    POSITIVE LOGITS
     help
    1.48
     allow
    1.40
     dictates
    1.36
     allowed
    1.32
     enabled
    1.30
     permits
    1.29
     already
    1.26
     permit
    1.22
     provide
    1.21
     enable
    1.18
    Activations Density 0.090%

    No Known Activations