© 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. 10-GEMMASCOPE-RES-131K
    4. 110888
    Prev
    Next
    INDEX
    Explanations

    woefully incomplete/incompetent

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

    The neuron strongly activates on longer content-bearing word fragments (i.e. non‐function subwords of moderate length), effectively spotting non‐stopword, multi‐character sequences.

    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
     through
    -2.02
     on
    -1.95
     some
    -1.84
     in
    -1.75
     with
    -1.73
    alties
    -1.70
     using
    -1.66
     not
    -1.63
     various
    -1.58
     such
    -1.54
    POSITIVE LOGITS
    และ
    1.53
     และ
    1.49
    }$,
    1.45
     многих
    1.38
    她的
    1.35
     являются
    1.34
     rapidamente
    1.30
     ſol
    1.30
     rapidement
    1.23
    而且
    1.21
    Activations Density 0.095%

    No Known Activations