© 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. 25679
    Prev
    Next
    INDEX
    Explanations

    seems to have been

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

    The neuron fires on copular and epistemic or passive‐voice verb forms (looks, seems, appears, is, have been discussed, etc.) that introduce statements or hedges.

    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
     przecież
    -1.50
     často
    -1.46
     często
    -1.24
     spesso
    -1.23
     zawsze
    -1.21
     oczywiście
    -1.16
     всегда
    -1.16
    adto
    -1.13
    jeuner
    -1.11
    ｇ
    -1.06
    POSITIVE LOGITS
     (
    1.20
     bastante
    1.17
     or
    1.17
    似乎
    1.14
     quite
    1.14
     verkar
    1.13
     /
    1.08
     only
    1.06
     ziemlich
    1.05
     '
    1.04
    Activations Density 0.107%

    No Known Activations