© 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. 22-GEMMASCOPE-RES-131K
    7. 130971
    Prev
    Next
    INDEX
    Explanations

    The main thing this neuron does is pick out strong first-person past-tense action verbs (e.g. “worked,” “studied,” “watched,” “ran”) signaling personal experience.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    past actions and states

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/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
     when
    -1.23
     or
    -1.15
     has
    -1.15
     after
    -1.15
     if
    -1.11
     now
    -1.08
     not
    -1.07
     perpetually
    -1.04
     new
    -1.02
     who
    -1.02
    POSITIVE LOGITS
    successfully
    1.30
    tried
    1.20
    
    1.18
    recently
    1.16
     leão
    1.12
    Tried
    1.10
    performed
    1.09
     enough
    1.09
    worked
    1.09
    played
    1.09
    Activations Density 0.105%

    No Known Activations