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. 3241
    Prev
    Next
    INDEX
    Explanations

    code structure elements and parameters in programming language syntax

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

    Code, coder, coding-related terms

    np_acts-logits-general · gemini-2.0-flash

    event codes and coding

    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
     Semp
    -0.63
    IUrlHelper
    -0.60
     Fant
    -0.59
     Fantasia
    -0.57
     Demp
    -0.55
     FANT
    -0.53
    InputTagHelper
    -0.51
    ParseTree
    -0.51
     يتيمه
    -0.49
    WTS
    -0.48
    POSITIVE LOGITS
     code
    2.00
    code
    1.86
     Code
    1.85
    Code
    1.85
     codes
    1.82
     coded
    1.72
     CODE
    1.72
    CODE
    1.69
     coding
    1.68
    coded
    1.65
    Activations Density 0.337%

    No Known Activations