© 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. 28162
    Prev
    Next
    INDEX
    Explanations

    mentioning names and titles

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

    The neuron strongly activates on tokens tied to sharing visual content—e.g. words like “Pic,” “share,” “took to Instagram,” “video”—i.e. mentions of posting or distributing photos/videos.

    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
     вашем
    -1.20
     ausgeschlossen
    -1.05
    strokeWeight
    -1.05
     setPassword
    -1.03
     or
    -1.01
     wtedy
    -0.99
     нашем
    -0.99
    presents
    -0.98
    性感
    -0.96
     xvi
    -0.96
    POSITIVE LOGITS
     ویتامین
    1.45
     ativar
    1.23
     jugu
    1.17
     atualizar
    1.15
     витами
    1.13
     alugar
    1.13
     коса
    1.12
     tatuagens
    1.12
     fizeram
    1.10
    unjukkan
    1.09
    Activations Density 0.024%

    No Known Activations