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. 32-GEMMASCOPE-RES-131K
    6. 130034
    Prev
    Next
    INDEX
    Explanations

    a mix of specific names and references, particularly related to individuals and places

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

    Chinese pinyin and names

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-9b-pt-res/layer_32/width_131k/average_l0_88
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    131,072
    Data Type
    float32
    Hook Name
    blocks.32.hook_resid_post
    Hook Layer
    32
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    :+:
    -0.43
    liferay
    -0.38
     blijven
    -0.37
    brum
    -0.37
     blijft
    -0.36
    Pyx
    -0.36
     dehydrated
    -0.36
    utilisons
    -0.36
     cause
    -0.35
     Cause
    -0.35
    POSITIVE LOGITS
    uo
    0.52
    ancang
    0.51
    édié
    0.49
    angling
    0.49
    ouren
    0.49
    AddTagHelper
    0.49
     keben
    0.47
    ao
    0.47
    Shi
    0.47
     étoient
    0.46
    Activations Density 0.073%

    No Known Activations