© 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
    3. Gemma-3-4B-IT
    5. 9-GEMMASCOPE-2-RES-16K
    7. 1621
    Prev
    Next
    INDEX
    Explanations

    negative consequences or continuations

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-4b-it/resid_post/layer_9_width_16k_l0_medium
    Prompts (Dashboard)
    238,145 prompts, 512 tokens each
    Dataset (Dashboard)
    lmsys + oasst1
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    <bos>
    1.40
     разделе
    1.35
    ка
    1.30
     become
    1.30
    те
    1.30
     مورد
    1.29
     provide
    1.26
    <em>
    1.25
    준
    1.23
    Со
    1.22
    POSITIVE LOGITS
     mismanagement
    1.83
     absurdity
    1.72
     reluctance
    1.67
     illusions
    1.65
     delusions
    1.62
     tweaking
    1.59
     Spannung
    1.58
     apari
    1.55
     cravings
    1.54
     objections
    1.53
    Activations Density 0.777%

    No Known Activations