Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-9B-IT
    4. Residual Stream - 131k
    5. 20-GEMMASCOPE-RES-131K
    6. 56500
    Prev
    Next
    INDEX
    Explanations

    keywords related to health monitoring and behavioral changes

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

    abnormal or suspicious states

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k
    Configuration
    google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    131,072
    Data Type
    float32
    Hook Name
    blocks.20.hook_resid_post
    Hook Layer
    20
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    interopRequire
    -0.51
    inflater
    -0.48
    vora
    -0.45
    anye
    -0.44
    ftagPool
    -0.43
     bast
    -0.43
     Pasteur
    -0.43
     engraçadas
    -0.42
    miyor
    -0.41
     насељу
    -0.41
    POSITIVE LOGITS
     sospe
    0.49
     abnormal
    0.42
    ötzlich
    0.41
     suspicious
    0.41
    tanleria
    0.38
     creș
    0.38
     FetchType
    0.37
     pinulongan
    0.36
     normalcy
    0.36
     anormal
    0.36
    Activations Density 0.085%

    No Known Activations