© 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. 466
    Prev
    Next
    INDEX
    Explanations

    TLR followed by numbers

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

    This neuron fires on structural or “meta” cues that mark new sections or headers—e.g. abstract‐style lead-ins (“Here we describe…”), email header fields (“cc:”, “ENRON”), and other document-level markers.

    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
     verkle
    -1.30
     essendo
    -1.21
    __;
    -1.20
    sekten
    -1.19
    \}\\
    -1.18
    middels
    -1.16
    凝土
    -1.15
    voer
    -1.15
    lecht
    -1.14
     verhind
    -1.13
    POSITIVE LOGITS
     konkrét
    1.14
     múl
    1.09
    cabo
    1.07
     Время
    1.02
    נוּ
    1.02
     rád
    1.01
    賄
    1.00
     inappropriés
    0.97
    颁
    0.94
    ग़
    0.93
    Activations Density 0.203%

    No Known Activations