© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Gemma-2-27B
    3. 10-GEMMASCOPE-RES-131K
    4. 67706
    Prev
    Next
    INDEX
    Explanations

    process of

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

    The neuron specializes in spotting common multi-word idiomatic or formulaic phrases—especially phrasal verbs and set expressions like “easy enough to,” “in the process of,” “far cry from,” “a whole host of,” 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_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
     когато
    -1.50
    詬
    -1.35
    してた
    -1.35
    面對
    -1.32
     esquecer
    -1.26
     กับ
    -1.25
     doporu
    -1.24
     here
    -1.21
     Wanneer
    -1.20
    ܵ
    -1.20
    POSITIVE LOGITS
     []:
    1.68
     coreana
    1.48
    。”
    1.45
     would
    1.43
    onné
    1.41
    ůli
    1.41
    return
    1.38
    西亚
    1.38
    CHREIBUNG
    1.33
    ();
    
    1.29
    Activations Density 0.091%

    No Known Activations