© 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. 34-GEMMASCOPE-RES-131K
    7. 93174
    Prev
    Next
    INDEX
    Explanations

    even followed by negation or minimality

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

    The neuron detects high-intensity degree adverbs and hedges—words like “even,” “remotely,” and “sure” that emphasize or qualify a statement.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
    quite
    -1.26
    だけでなく
    -1.17
     also
    -1.13
     également
    -1.09
     indeed
    -1.09
     菇
    -1.05
     spektak
    -1.02
     TAMBÉM
    -1.02
     CERTAIN
    -1.02
    indeed
    -1.02
    POSITIVE LOGITS
     attempted
    1.53
     remotely
    1.46
     siquiera
    1.45
     attempt
    1.38
     begin
    1.27
     semblance
    1.27
     vaguely
    1.26
     close
    1.20
     hinted
    1.20
     bother
    1.18
    Activations Density 0.029%

    No Known Activations