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-2-9B
    4. Residual Stream - 131k
    5. 20-GEMMASCOPE-RES-131K
    6. 70685
    Prev
    Next
    INDEX
    Explanations

    elements related to time and numerical references

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

    Characters or short sequences that form part of structured text elements like citations, mathematical notation, dates, file paths, or reference markers in academic or technical documents.

    eleuther_acts_top20 · claude-3-7-sonnet-20250219Triggered by @lu-christina
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-9b-pt-res/layer_20/width_131k/average_l0_114
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    131,072
    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
    bitField
    -0.62
    aarrggbb
    -0.60
    writeFieldEnd
    -0.57
     мәкал
    -0.52
    GEBURTSDATUM
    -0.51
    thâu
    -0.51
     resourceCulture
    -0.50
     Vergnügen
    -0.48
     télévision
    -0.48
    Aholisi
    -0.47
    POSITIVE LOGITS
    setVerticalGroup
    0.49
    合
    0.42
    spill
    0.41
    StoryboardSegue
    0.39
    قص
    0.38
    انيف
    0.38
     surla
    0.36
     nakalista
    0.36
     kina
    0.36
    JsonPropertyName
    0.36
    Activations Density 0.075%

    No Known Activations