© 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. 10-GEMMASCOPE-RES-131K
    4. 111057
    Prev
    Next
    INDEX
    Explanations

    state of affairs

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

    The neuron detects verbs (often in phrasal or stative uses like “getting serious,” “pick up,” “get strange,” “looking up,” “got better”) that signal a shift or change in the state of affairs.

    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
     stále
    -1.42
     sorprender
    -1.41
    ")]
    -1.41
    qtt
    -1.38
     ´
    -1.34
    "));
    -1.33
     calientes
    -1.31
     salpic
    -1.30
    去
    -1.29
    ensões
    -1.28
    POSITIVE LOGITS
     only
    1.66
    あっという間に
    1.46
    ほうがいい
    1.41
    そこから
    1.32
    方がいい
    1.31
    互動
    1.26
     vérité
    1.25
     immediately
    1.25
    欣賞
    1.24
     pukul
    1.21
    Activations Density 0.025%

    No Known Activations