© 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. 75042
    Prev
    Next
    INDEX
    Explanations

    startling reversal, observation, reactions

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

    The neuron consistently fires strongest on uncommon, content-bearing words (e.g. technical terms or otherwise rare vocabulary) rather than on common function words.

    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
     While
    -2.55
     Upon
    -2.45
     Thus
    -2.45
     Once
    -2.41
     If
    -2.34
     Therefore
    -2.28
     has
    -2.25
    </td>
    -2.23
     Meanwhile
    -2.20
    Ꮢ
    -2.20
    POSITIVE LOGITS
    e
    3.55
    齪
    3.06
    牁
    3.03
    皤
    2.81
    </strong>
    2.58
    휀
    2.58
    {
    2.44
    珰
    2.42
     problemy
    2.38
    ୮
    2.33
    Activations Density 0.001%

    No Known Activations