© 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. 10-GEMMASCOPE-RES-131K
    7. 90468
    Prev
    Next
    INDEX
    Explanations

    sunken shipwreck or ship

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

    The neuron fires on unusually long contiguous tokens—e.g. long class, method, domain, or other multi‐syllable identifiers.

    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
    ER
    -2.53
    ⫺
    -2.31
    said
    -2.28
    ES
    -2.25
    最重要的
    -2.16
    自己的
    -2.14
    EM
    -2.13
    IN
    -2.11
     femenino
    -2.11
    特殊的
    -2.09
    POSITIVE LOGITS
     conocen
    2.45
    姈
    2.41
     Einladung
    2.38
    ৢ
    2.34
    しないで
    2.27
     Gegenstände
    2.25
     Getränke
    2.25
     Which
    2.25
    狆
    2.19
     Ergänzung
    2.17
    Activations Density 0.001%

    No Known Activations