© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-2B
    4. Residual Stream - 16k
    5. 20-GEMMASCOPE-RES-16K
    6. 8335
    Prev
    Next
    INDEX
    Explanations

    mentions of geographical locations and associated environmental contexts

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

    Islands and islanders

    np_acts-logits-general · gemini-2.0-flash

    island locations and island nations

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 20-gemmascope-res-16k
    Configuration
    google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    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
     CERN
    -0.53
    elstein
    -0.52
     Cheltenham
    -0.48
     Rye
    -0.48
    wechat
    -0.47
     Iberia
    -0.47
     Blackstone
    -0.47
    Rye
    -0.47
    يكا
    -0.47
     Iberian
    -0.46
    POSITIVE LOGITS
     Islands
    1.32
     ISLANDS
    1.23
     islands
    1.21
     island
    1.17
    Islands
    1.15
     Island
    1.10
     islanders
    1.06
     Islander
    1.06
    islands
    1.05
    island
    1.04
    Activations Density 0.142%

    No Known Activations