© 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. 98695
    Prev
    Next
    INDEX
    Explanations

    methods and improvements

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

    The neuron fires strongly on verbs and nouns that propose solutions or describe methods—i.e. the key action‐and‐method words in technical or how-to passages.

    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
     iniciativas
    -1.12
    andangan
    -1.06
     análisis
    -0.97
     diseñ
    -0.95
    できなかった
    -0.94
     fábricas
    -0.93
     cozin
    -0.93
     indicaciones
    -0.91
    なのですが
    -0.89
    でしたか
    -0.89
    POSITIVE LOGITS
    much
    1.21
     instead
    1.11
     lada
    1.09
    不但
    1.05
    both
    1.05
     better
    1.04
    better
    1.03
     TOD
    1.02
     SENT
    1.01
     гораздо
    1.01
    Activations Density 0.147%

    No Known Activations