© 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. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    5. Gemma-2-9B-IT
    7. Residual Stream - 131k
    9. 31-GEMMASCOPE-RES-131K
    11. 23664
    Prev
    Next
    INDEX
    Explanations

    past tense forms of verbs related to events or actions involving people

    oai_token-act-pair · gpt-4o-miniTriggered by @bot

    arrested, killed, or found

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B-IT @ 31-gemmascope-res-131k
    Configuration
    google/gemma-scope-9b-it-res/layer_31/width_131k/average_l0_109
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    131,072
    Data Type
    float32
    Hook Name
    blocks.31.hook_resid_post
    Hook Layer
    31
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     varones
    -0.43
     Feind
    -0.39
     împre
    -0.38
     CreateTagHelper
    -0.37
     ohnehin
    -0.37
     sebelumnya
    -0.36
     vorher
    -0.35
     zuvor
    -0.35
     Hinblick
    -0.35
     Schließlich
    -0.35
    POSITIVE LOGITS
     arrested
    1.09
     arrest
    0.95
     Arrest
    0.94
     killed
    0.87
    Arrest
    0.86
     Arrested
    0.86
    arrest
    0.85
     sentenced
    0.85
     jailed
    0.80
     murdered
    0.77
    Activations Density 0.066%

    No Known Activations