© 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. 22-GEMMASCOPE-RES-131K
    7. 25801
    Prev
    Next
    INDEX
    Explanations

    conjunctions of consequence

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

    The neuron strongly activates on semantically rich “content” words (nouns, verbs, adjectives, numbers) and remains inactive on common function words.

    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
     plateado
    -1.02
     atractivo
    -1.01
     valioso
    -0.98
     hidratante
    -0.95
     decorado
    -0.94
    もありますが
    -0.92
     houten
    -0.90
    ませんでした
    -0.90
    StatusLabel
    -0.89
     élégante
    -0.87
    POSITIVE LOGITS
     so
    4.03
     поэтому
    2.98
    所以
    2.75
     nên
    2.38
     所以
    2.23
    于是
    2.17
     therefore
    2.16
     więc
    2.14
     So
    2.09
    so
    2.06
    Activations Density 0.148%

    No Known Activations