© 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. 10-GEMMASCOPE-RES-131K
    7. 21019
    Prev
    Next
    INDEX
    Explanations

    version numbers and technical terms

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

    The neuron is picking out configuration or template field names—i.e. the keys (often ending in “:”) in structured config/JSON/YAML snippets.

    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
     universelle
    -1.51
     AMAZING
    -1.50
    𝑈
    -1.49
     alarme
    -1.40
     VERY
    -1.39
    ἵ
    -1.38
     spirituelle
    -1.38
     rigide
    -1.37
     kokos
    -1.30
     superstar
    -1.29
    POSITIVE LOGITS
     such
    1.57
     verbeter
    1.46
     at
    1.43
     vertrou
    1.36
     gevoel
    1.32
     their
    1.29
     herinner
    1.25
     twij
    1.22
     considerable
    1.21
     quidem
    1.21
    Activations Density 0.027%

    No Known Activations