© 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. 22-GEMMASCOPE-RES-131K
    4. 70776
    Prev
    Next
    INDEX
    Explanations

    negation

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

    The neuron detects logical negation markers (e.g. “!” or “not”) in code.

    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
    かなかった
    -1.19
     différentes
    -1.16
    
    -1.09
     légèrement
    -1.07
     différents
    -1.07
     prachtig
    -1.05
     частью
    -1.05
    囔
    -1.03
    extremely
    -1.02
    slightly
    -1.02
    POSITIVE LOGITS
    含量
    1.00
     wykonane
    0.99
     engaging
    0.99
    !!!!!
    0.96
     If
    0.94
    viert
    0.94
    It
    0.93
    ײ
    0.91
     k
    0.91
    胜
    0.89
    Activations Density 0.007%

    No Known Activations