© 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. 34-GEMMASCOPE-RES-131K
    7. 117316
    Prev
    Next
    INDEX
    Explanations

    according to a report

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

    The neuron fires strongly on most of the main content words—especially numbers, verbs, and nouns—in the opening sentences (the “lede”) of news‐style articles.

    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
     aforementioned
    -1.34
    Adapun
    -1.13
    连忙
    -1.09
     takich
    -1.08
    ributed
    -1.05
     other
    -1.05
    ává
    -1.05
    genommen
    -1.04
     différents
    -1.03
    らかに
    -1.03
    POSITIVE LOGITS
    这里
    1.29
    ംബ
    1.10
    1.07
    ropshire
    1.03
     newly
    1.00
    Demografie
    0.97
     offenbar
    0.97
    将会
    0.95
     erneut
    0.94
    uju
    0.94
    Activations Density 0.580%
    No Known Activations