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-IT
    4. Residual Stream - 16k
    5. 31-GEMMASCOPE-RES-16K
    6. 16234
    Prev
    Next
    INDEX
    Explanations

    data structures and manipulation operations in programming code

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

    Code that contains lists or arrays

    np_acts-logits-general · gemini-2.0-flash

    code constructs and data structures

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B-IT @ 31-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-it-res/layer_31/width_16k/average_l0_76
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.31.hook_resid_post
    Hook Layer
    31
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     skul
    -0.27
     tani
    -0.25
     over
    -0.25
     Konzept
    -0.25
    subplots
    -0.25
    ству
    -0.25
    Queues
    -0.24
     kati
    -0.24
    timeScale
    -0.24
     grants
    -0.23
    POSITIVE LOGITS
     betweenstory
    0.87
    parsedMessage
    0.72
    rungsseite
    0.69
     kasarigan
    0.69
    期刊论文
    0.67
    ⓧ
    0.66
     AssemblyTitle
    0.65
     Walkover
    0.62
     Italijanski
    0.60
     utafitiHapana
    0.60
    Activations Density 0.154%

    No Known Activations