© 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. 33222
    Prev
    Next
    INDEX
    Explanations

    proposing or arguing

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

    The neuron detects formal legal or academic report language—especially words tied to court proceedings and procedures (e.g. case, evidence, trial, should receive).

    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
    Acknowledgement
    -0.96
    ADOR
    -0.93
    ACKNOWLEDGEMENTS
    -0.90
    tré
    -0.88
    ⁂
    -0.85
    ăsp
    -0.85
    ʁ
    -0.82
    []=
    -0.82
     appréci
    -0.82
     discrepan
    -0.81
    POSITIVE LOGITS
     proposed
    1.66
     justify
    1.20
     argument
    1.20
     argued
    1.20
     proposal
    1.17
     Proposed
    1.14
     argues
    1.12
     propuesto
    1.11
     arguments
    1.10
    Justification
    1.09
    Activations Density 0.078%

    No Known Activations