© 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. 84559
    Prev
    Next
    INDEX
    Explanations

    scores and results

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

    The neuron is detecting perfect‐aspect verb constructions (e.g. “has beaten,” “have been,” “including,” 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_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
    ">!
    -0.98
    の中から
    -0.96
     peduli
    -0.96
    iction
    -0.96
    issipp
    -0.95
     ayude
    -0.94
     whoſe
    -0.94
     élevés
    -0.92
     between
    -0.92
     registre
    -0.92
    POSITIVE LOGITS
     twice
    1.21
    最近
    1.18
     lately
    1.13
     recently
    1.09
    Twice
    1.05
     récemment
    1.05
    agliari
    1.01
     recientemente
    0.99
     March
    0.97
    هنگ
    0.97
    Activations Density 0.017%

    No Known Activations