© 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
    2. Gemma-2-27B
    3. 22-GEMMASCOPE-RES-131K
    4. 67383
    Prev
    Next
    INDEX
    Explanations

    create, store, update actions

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

    The neuron strongly activates on programming action names—verbs like “create,” “update,” “check,” “input,” etc.—i.e. tokens naming controller actions or request‐handling endpoints.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/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
     this
    -0.97
     convoc
    -0.91
    thia
    -0.91
     turísticos
    -0.91
    izações
    -0.91
     more
    -0.88
     Magen
    -0.88
    憊
    -0.87
     putri
    -0.85
    Boarding
    -0.85
    POSITIVE LOGITS
     submits
    0.98
     submission
    0.98
     submit
    0.94
    PostMapping
    0.88
     submitted
    0.85
     outcome
    0.82
     finalizar
    0.82
    ℏ
    0.79
     then
    0.79
     onSubmit
    0.79
    Activations Density 0.005%

    No Known Activations