© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 25235
    Prev
    Next
    INDEX
    Explanations

    safety and efficacy

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

    The neuron spotlights terms denoting study or performance outcomes (e.g. efficacy, safety, impact, success, distress, avoid).

    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
    avice
    -1.11
     Процитовано
    -1.08
    歺
    -1.06
    灕
    -1.03
     quatrième
    -1.01
     herunterladen
    -1.01
     strengthens
    -0.99
     surpasses
    -0.99
     vierte
    -0.99
    khó
    -0.98
    POSITIVE LOGITS
     safety
    1.17
    V
    1.07
     security
    1.06
     cost
    1.05
     legal
    1.02
     productivity
    1.00
     democratic
    1.00
     economic
    0.98
     slow
    0.96
     laws
    0.96
    Activations Density 0.078%

    No Known Activations