© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 13943
    Prev
    Next
    INDEX
    Explanations

    always be

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

    The neuron strongly activates on words that express personal regret or unfulfilled desire—“I wish I could’ve saved more,” “never enough time,” etc.

    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
     информация
    -0.84
    achite
    -0.82
    ciuto
    -0.81
    其它
    -0.80
    我来
    -0.78
     menghindari
    -0.77
    Revenir
    -0.77
    ROIT
    -0.74
     eccentricity
    -0.73
    SUBST
    -0.73
    POSITIVE LOGITS
     always
    1.43
     תמיד
    1.26
     zawsze
    1.26
     всегда
    1.24
     alltid
    1.23
     mindig
    1.21
     toujours
    1.20
     sempre
    1.15
    ALWAYS
    1.15
    Always
    1.12
    Activations Density 0.047%

    No Known Activations