© 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. 10-GEMMASCOPE-RES-131K
    7. 12223
    Prev
    Next
    INDEX
    Explanations

    erase, erasing, erasure

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

    The neuron activates on occurrences of “erase” (and its morphological variants like erasing, erasure, erasable).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
    IN
    -2.08
    s
    -1.96
     mauva
    -1.89
    C
    -1.87
    P
    -1.84
     he
    -1.84
    饪
    -1.83
    츨
    -1.82
    므
    -1.82
    IS
    -1.81
    POSITIVE LOGITS
     llegamos
    2.02
     sillä
    1.95
     fortalec
    1.90
    ла
    1.90
     kült
    1.89
     sabotage
    1.84
     partnering
    1.84
    oncesto
    1.84
     estábamos
    1.83
    berfläche
    1.80
    Activations Density 0.007%

    No Known Activations