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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & 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. 12675
    Prev
    Next
    INDEX
    Explanations

    concepts related to structure and organization in both physical and abstract contexts

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

    Words/phrases followed by relative pronouns

    np_acts-logits-general · gemini-2.0-flash

    words followed by relative clauses

    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
    ArrowToggle
    -0.80
    CppCodeGen
    -0.78
    IVEREF
    -0.78
    aarrggbb
    -0.74
    bootstrapcdn
    -0.69
     AssemblyCulture
    -0.68
    AddTagHelper
    -0.65
    ownic
    -0.63
    ::*;
    -0.63
    DotNetBar
    -0.60
    POSITIVE LOGITS
     that
    2.50
     which
    2.09
     who
    1.67
     whose
    1.67
    which
    1.50
     که
    1.49
    that
    1.44
     которые
    1.36
    whose
    1.32
     που
    1.31
    Activations Density 3.420%

    No Known Activations