© 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
    2. Gemma-2-27B
    3. 10-GEMMASCOPE-RES-131K
    4. 75471
    Prev
    Next
    INDEX
    Explanations

    from there

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

    The neuron is sensitive to deictic transition markers—adverbs and phrases that signal a shift or progression in time, place, or discourse (e.g. from here on out, at that point).

    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
     TWO
    -1.42
    ܜ
    -1.41
    делы
    -1.41
    壜
    -1.37
    ╫
    -1.37
     ALWAYS
    -1.37
    ThemeOverlay
    -1.34
    蜕
    -1.32
     regresa
    -1.31
     skrevet
    -1.31
    POSITIVE LOGITS
     here
    1.66
    就好
    1.57
    を入力
    1.47
     beyond
    1.45
     me
    1.39
    here
    1.39
     divent
    1.38
     thence
    1.34
    ly
    1.32
     genera
    1.31
    Activations Density 0.017%

    No Known Activations