© 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. 9380
    Prev
    Next
    INDEX
    Explanations

    Reuters datelines

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

    The neuron strongly activates on the little header/meta‐data bits of a Reuters newswire story—things like the “(Reuters)” source tag, dates, locations, and other numeric tokens.

    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
     otel
    -1.11
     hura
    -1.11
     gitar
    -1.10
     hien
    -1.10
     karbon
    -1.09
     Amerik
    -1.09
    ſa
    -1.08
     himo
    -1.06
     traktor
    -1.05
     kupa
    -1.05
    POSITIVE LOGITS
     provved
    1.03
     Bewußt
    1.00
     gäng
    1.00
     gött
    0.99
     hust
    0.97
    🥽
    0.97
    𝒱
    0.97
     behandelt
    0.95
     gän
    0.93
    ările
    0.92
    Activations Density 0.004%

    No Known Activations