© 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. 35848
    Prev
    Next
    INDEX
    Explanations

    temporal markers like when, before, until

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

    The neuron detects second‐person cautionary advice or warnings phrased as conditional statements (e.g. “if you wait… you may…”).

    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
     quaisquer
    -1.21
    積極的に
    -1.09
    が広
    -1.07
    
    -1.06
    夜は
    -1.05
    properly
    -1.02
     douze
    -1.02
     fatores
    -1.01
    contrar
    -0.99
     pensamentos
    -0.99
    POSITIVE LOGITS
     when
    1.38
     cuando
    1.25
     or
    1.21
     before
    1.16
     it
    1.13
     When
    1.12
     уже
    1.12
     потім
    1.09
     first
    1.09
     becomes
    1.08
    Activations Density 0.023%

    No Known Activations