© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 45722
    Prev
    Next
    INDEX
    Explanations

    improved or enhanced states

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

    The neuron detects past‐participle verbs ending in “-ed” that denote a change or adjustment (e.g. improved, reduced, enhanced, increased, modified).

    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
    และ
    -1.33
     Aunque
    -1.32
    𓁹
    -1.28
     wacky
    -1.27
    despite
    -1.26
     actuellement
    -1.26
    馓
    -1.23
    ⚥
    -1.23
    regardless
    -1.23
     différents
    -1.23
    POSITIVE LOGITS
     version
    2.02
     versions
    1.60
     from
    1.35
     focused
    1.35
     used
    1.31
     and
    1.31
    version
    1.24
     helped
    1.22
     focusing
    1.21
    After
    1.20
    Activations Density 0.051%

    No Known Activations