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-2-2B
    4. Residual Stream - 16k
    5. 20-GEMMASCOPE-RES-16K
    6. 13487
    Prev
    Next
    INDEX
    Explanations

    patterns of numerical or mathematical data

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

    "3" or "three"

    np_acts-logits-general · gemini-2.0-flash

    three

    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
     useStyles
    -0.58
     Tenth
    -0.52
     contextLoads
    -0.51
    interopRequire
    -0.51
    utnik
    -0.50
    oneofs
    -0.48
    󠁢
    -0.48
     tenth
    -0.47
    SBATCH
    -0.47
     lunedì
    -0.47
    POSITIVE LOGITS
    3
    1.55
     Third
    1.32
     third
    1.26
     ketiga
    1.18
     trio
    1.18
     three
    1.16
    Third
    1.16
     Three
    1.16
    3
    1.15
     thirds
    1.15
    Activations Density 1.352%

    No Known Activations