© 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. 34-GEMMASCOPE-RES-131K
    4. 2359
    Prev
    Next
    INDEX
    Explanations

    prepositional phrases initiating descriptions

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

    The neuron tends to fire on semantically salient content words—uncommon nouns and verbs that carry the main topical or thematic meaning of the text.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
     ܝ
    -1.20
    andaag
    -1.06
     oggi
    -1.06
     heiligen
    -1.05
     frattempo
    -1.04
     Heilige
    -0.97
     dabar
    -0.96
     Oggi
    -0.91
     heilig
    -0.91
    فقط
    -0.91
    POSITIVE LOGITS
     from
    5.75
     right
    4.56
     dès
    4.34
     desde
    4.13
     vanaf
    3.34
     từ
    3.20
    从
    3.00
    right
    2.97
     ngay
    2.95
    부터
    2.94
    Activations Density 0.106%

    No Known Activations