© 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. 10-GEMMASCOPE-RES-131K
    7. 43912
    Prev
    Next
    INDEX
    Explanations

    future state after condition

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

    The neuron strongly activates on occurrences of the word “once,” i.e. it detects that temporal‐sequence cue.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
     was
    -1.83
     “
    -1.83
     -
    -1.70
     accompanying
    -1.63
    4
    -1.61
     -}
    -1.60
    0
    -1.59
     --
    -1.56
    ("
    -1.55
     ...
    -1.55
    POSITIVE LOGITS
     vulné
    2.03
     fueran
    1.72
    Once
    1.63
    
    1.63
     Sobald
    1.59
    員工
    1.57
    　　「
    1.57
    sandália
    1.56
    beitung
    1.55
    以上です
    1.53
    Activations Density 0.024%

    No Known Activations