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-2-9B-IT
    4. Residual Stream - 16k
    5. 20-GEMMASCOPE-RES-16K
    6. 9781
    Prev
    Next
    INDEX
    Explanations

    words and phrases related to accuracy, precision, or reliability in information or measurements.

    oai_token-act-pair · claude-3-5-sonnet-20240620Triggered by @myinzi123

    words and phrases related to accuracy and reliability in contexts like evaluation and diagnosis

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

    Adverbs modifying accuracy, reliability, and safety

    np_acts-logits-general · gemini-2.0-flash

    accuracy and correctness

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91
    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
     utafitiHapana
    -0.36
    เอ
    -0.32
     apart
    -0.31
     EP
    -0.31
     Feld
    -0.31
     FIELDS
    -0.31
     Fleisch
    -0.31
     dumb
    -0.31
     bat
    -0.31
     fields
    -0.30
    POSITIVE LOGITS
    Datuak
    0.71
    ftagPool
    0.71
    Accurate
    0.71
    bootstrapcdn
    0.70
     Accurate
    0.69
    Proper
    0.66
    декват
    0.66
    Reasonable
    0.65
    Appropriate
    0.64
     Adequate
    0.64
    Activations Density 0.747%

    No Known Activations