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 - 131k
    5. 20-GEMMASCOPE-RES-131K
    6. 45733
    Prev
    Next
    INDEX
    Explanations

    phrases related to qualifications and strengths in various contexts

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

    concessive conjunctions and their equivalents

    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-131k
    Configuration
    google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81
    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
    Erreferentziak
    -0.43
     ویکی‌پدی
    -0.39
     resourceCulture
    -0.38
     timing
    -0.37
    الدراسه
    -0.37
    tanleria
    -0.36
    Atsauces
    -0.35
     мәкалә
    -0.35
     confuse
    -0.35
     PARSER
    -0.35
    POSITIVE LOGITS
     nevertheless
    0.77
     nonetheless
    0.73
    theless
    0.64
     ändå
    0.60
    etheless
    0.59
    Nevertheless
    0.58
    それでも
    0.58
     Nevertheless
    0.57
    Nonetheless
    0.53
     néanmoins
    0.52
    Activations Density 0.045%

    No Known Activations