© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 79496
    Prev
    Next
    INDEX
    Explanations

    sound duration and sustain

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

    The neuron fires on technical audio/music terms—e.g. “sample,” “note,” “stroke,” “Stop All,” “MIDI,” etc.—marking domain‐specific sound/timing vocabulary.

    oai_token-act-pair · o4-mini
    Triggered 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
    Negative Logits
     ken
    -0.77
    tempo
    -0.75
    bolite
    -0.73
    odom
    -0.73
     PACE
    -0.72
    Dissolved
    -0.72
    בודה
    -0.71
     CBE
    -0.69
    projeto
    -0.68
     mated
    -0.68
    POSITIVE LOGITS
     sustain
    1.38
     Sustain
    1.27
     played
    1.16
     sustains
    1.10
    Sustain
    1.10
     duration
    1.08
     envelopes
    1.08
     volume
    1.08
     sustaining
    1.05
     ADS
    1.04
    Activations Density 0.009%

    No Known Activations