© 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. 80601
    Prev
    Next
    INDEX
    Explanations

    -*- coding: utf-8 -*-

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

    The neuron is picking out structural or formatting tokens—especially punctuation like commas in dateline headers and encoding‐comment markers—rather than ordinary words.

    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
     persoane
    -1.44
     into
    -1.37
     gouverneur
    -1.31
    ☫
    -1.27
    薷
    -1.23
     femei
    -1.23
    razioni
    -1.22
    NATION
    -1.22
     pensent
    -1.21
    ränk
    -1.21
    POSITIVE LOGITS
    !
    1.50
    ,”
    1.40
     gdzie
    1.29
     where
    1.27
    。
    1.27
     их
    1.25
     was
    1.23
     musia
    1.23
    basicConfig
    1.20
    ”（
    1.19
    Activations Density 0.272%

    No Known Activations