© 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
    2. Gemma-2-27B
    3. 22-GEMMASCOPE-RES-131K
    4. 40167
    Prev
    Next
    INDEX
    Explanations

    locations in countries

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

    Explanation of neuron 4 behavior: the main thing this neuron does is find prepositions that introduce spatial or temporal relations (e.g. in, on, to, outside, prior).

    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
     independente
    -1.25
    少了
    -1.25
    只是
    -1.15
     dessa
    -1.13
     بدأ
    -1.09
     lungo
    -1.08
    <bos>
    -1.07
    hline
    -1.05
    ▃
    -1.04
     adegu
    -1.03
    POSITIVE LOGITS
     or
    1.83
     country
    1.29
     mengapa
    1.26
     Australia
    1.25
     countries
    1.23
    咔嚓
    1.23
     within
    1.13
     different
    1.13
     their
    1.13
     areas
    1.10
    Activations Density 0.023%

    No Known Activations