© 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. 22-GEMMASCOPE-RES-131K
    7. 16056
    Prev
    Next
    INDEX
    Explanations

    believe that should

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

    This neuron detects expressions of belief or opinion, especially instances of “believe that” (and similar stance-taking verbs plus “that”).

    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
     чехол
    -0.89
    ágot
    -0.88
    bundet
    -0.88
    ará
    -0.86
    blower
    -0.85
    NetworkInfo
    -0.84
    kirakan
    -0.84
     zda
    -0.84
    antaranya
    -0.82
    MÁS
    -0.82
    POSITIVE LOGITS
    应该
    1.51
     should
    1.48
     должны
    1.30
     shouldn
    1.26
    應該
    1.21
     должна
    1.19
     πρέπει
    1.13
     Should
    1.13
     sollte
    1.12
     צריך
    1.07
    Activations Density 0.063%

    No Known Activations