© 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. 73415
    Prev
    Next
    INDEX
    Explanations

    temporal gap or consequence

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

    The neuron detects temporal “after” phrases that include numeric references to elapsed time (e.g., “after 302,” “after 300,” 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_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
    enablog
    -1.63
     are
    -1.59
     crumbling
    -1.59
    batore
    -1.53
     ferner
    -1.52
     roads
    -1.50
    }
    -1.49
    出了一
    -1.48
    Namun
    -1.48
     time
    -1.45
    POSITIVE LOGITS
     its
    1.98
     it
    1.94
     maksi
    1.64
     that
    1.63
     when
    1.62
     définiti
    1.56
    ledes
    1.55
    槼
    1.52
     septem
    1.51
     hakk
    1.47
    Activations Density 0.022%

    No Known Activations