© 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. 103660
    Prev
    Next
    INDEX
    Explanations

    permanence versus deletion

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

    The main thing this neuron does is detect terms and phrases related to deleting or removing items (e.g. “delete,” “remove,” “rows deleted,” etc.).

    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
    ₧
    -0.99
     ),
    -0.91
    ancias
    -0.87
    Inser
    -0.85
    ADD
    -0.85
     installing
    -0.84
    InsertCommand
    -0.83
    INSERT
    -0.82
    /**
    -0.82
     creating
    -0.82
    POSITIVE LOGITS
    ById
    1.37
     forever
    1.23
    forever
    1.06
    perman
    1.05
    Perman
    1.05
     vĩnh
    1.04
     Perman
    1.03
     permanently
    1.03
     perman
    0.96
     навсегда
    0.95
    Activations Density 0.003%

    No Known Activations