© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 66484
    Prev
    Next
    INDEX
    Explanations

    Herald and related words

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

    This neuron detects newspaper source mentions, especially instances of “Herald” in bylines or citations.

    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
    塍
    -2.95
     ComVisible
    -2.47
     tomado
    -2.44
    熺
    -2.42
     "...
    -2.28
    。"
    -2.27
     是否
    -2.23
    鋱
    -2.23
    呦
    -2.23
     atau
    -2.22
    POSITIVE LOGITS
    to
    3.34
    “
    2.69
    from
    2.48
    they
    2.45
    ”:
    2.42
    是
    2.42
    ,“
    2.34
    :“
    2.30
     Then
    2.28
     ,‬
    2.27
    Activations Density 0.002%

    No Known Activations