© 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. 69350
    Prev
    Next
    INDEX
    Explanations

    ways to

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

    The neuron fires strongly on the word “to” when it appears in list-style headings (e.g. “Reasons to…,” “Ways to…,” “Things to…”)

    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
    -¡
    -1.59
     ausein
    -1.56
    】:
    -1.54
     gjorde
    -1.53
     TType
    -1.52
    yarakat
    -1.51
    iritto
    -1.49
    太过
    -1.46
     exped
    -1.43
     ダイ
    -1.41
    POSITIVE LOGITS
     neue
    1.57
    :
    1.41
     new
    1.40
     nosso
    1.38
    ?
    1.38
     nimi
    1.36
     not
    1.35
    这么多
    1.34
    ᯾
    1.33
    萏
    1.32
    Activations Density 0.012%

    No Known Activations