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. 20-GEMMASCOPE-RES-16K
    6. 12856
    Prev
    Next
    INDEX
    Explanations

    locations or references to 'here' in various contexts

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    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
    granate
    -0.51
     опро
    -0.51
    Magento
    -0.50
     Ethiopian
    -0.50
     filt
    -0.49
     Activision
    -0.49
    UOUS
    -0.49
     Tibetan
    -0.48
     myſelf
    -0.48
     Caspian
    -0.48
    POSITIVE LOGITS
    here
    1.20
     here
    1.15
     Here
    1.11
    Here
    1.10
     HERE
    1.06
    HERE
    1.05
     aquí
    0.97
     Aquí
    0.93
     tää
    0.90
     здесь
    0.87
    Activations Density 0.095%

    No Known Activations