© 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. 34-GEMMASCOPE-RES-131K
    7. 59914
    Prev
    Next
    INDEX
    Explanations

    common phrases

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

    The neuron fires on (mostly high‐value) verbs and modifiers that describe changing or adjusting something—words like “give,” “take,” “add,” “clear,” 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_34/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
    Dinosaur
    -0.74
     therefrom
    -0.72
    demie
    -0.72
     acess
    -0.71
    入手
    -0.70
     прид
    -0.70
    ↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵↵
    -0.69
     anew
    -0.69
    constra
    -0.69
     Tait
    -0.69
    POSITIVE LOGITS
     things
    0.85
     stuff
    0.84
     التاريخ
    0.80
     org
    0.79
    syn
    0.76
     البلد
    0.75
    东西
    0.75
    東西
    0.75
     einiges
    0.73
     troublesome
    0.73
    Activations Density 0.051%

    No Known Activations