© 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. 58556
    Prev
    Next
    INDEX
    Explanations

    too or not bad

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

    The neuron is triggered by degree words and negation (adverbs like “too,” “not,” “half,” etc.) that modify or intensify statements.

    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
    íncia
    -0.85
    īm
    -0.82
     Décembre
    -0.82
    ousness
    -0.73
    ξης
    -0.73
     quartered
    -0.72
    шев
    -0.71
    ania
    -0.71
     abomination
    -0.71
     Strengths
    -0.69
    POSITIVE LOGITS
     bad
    6.81
    bad
    5.13
    Bad
    4.72
     Bad
    4.47
     BAD
    3.72
    BAD
    3.66
    坏
    2.69
     بد
    2.55
     kötü
    2.34
     плохо
    2.30
    Activations Density 0.025%

    No Known Activations