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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-9B
    4. Residual Stream - 16k
    5. 20-GEMMASCOPE-RES-16K
    6. 11331
    Prev
    Next
    INDEX
    Explanations

    programming constructs related to mathematical functions and calculations

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

    mathematical comparison operations, particularly "max" functions in code and comparative phrases like "latest of" or "greater of".

    oai_token-act-pair · claude-3-7-sonnet-20250219Triggered by @neilrathi

    comparative or superlative constructions—especially in text or code—that select the maximum/greater value among alternatives or indicate increasing upper bounds.

    oai_token-act-pair · gpt-5Triggered by @jyhe0408

    the word "max" in the context of mathematical operations or comparisons, particularly when followed by an opening parenthesis.

    oai_token-act-pair · claude-4-5-sonnetTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B @ 20-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-pt-res/layer_20/width_16k/average_l0_68
    Prompts (Dashboard)
    24,576 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
     aras
    -0.38
     муніципалі
    -0.36
    "])
    
    -0.34
    节
    -0.34
    BASEPATH
    -0.34
     Fee
    -0.33
     rękaw
    -0.33
     dựng
    -0.33
     door
    -0.33
     cons
    -0.33
    POSITIVE LOGITS
     MainAxisSize
    0.63
    oredCriteria
    0.62
    tagHelperRunner
    0.54
     ब्रेकडाउन
    0.54
    orElse
    0.50
    onedDateTime
    0.48
    webElementXpaths
    0.47
     gynhyrchwyd
    0.46
    windowFixed
    0.45
    MessageTagHelper
    0.45
    Activations Density 0.244%

    No Known Activations