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-2-2B
    4. Residual Stream - 16k
    5. 20-GEMMASCOPE-RES-16K
    6. 4672
    Prev
    Next
    INDEX
    Explanations

    phrases relating to performance and evaluation outcomes

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

    high performance

    np_acts-logits-general · gemini-2.0-flash

    performance quality

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 20-gemmascope-res-16k
    Configuration
    google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71
    Prompts (Dashboard)
    36,864 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
    Shortest
    -0.44
     rodada
    -0.43
     了
    -0.42
    Necessary
    -0.41
    direct
    -0.40
    WebMethod
    -0.40
     Lasting
    -0.39
    CppCodeGen
    -0.39
     tabi
    -0.39
    Kanpo
    -0.38
    POSITIVE LOGITS
     well
    1.25
     poorly
    1.23
     excellently
    1.08
     better
    1.05
     favourably
    1.02
     badly
    1.02
     favorably
    0.99
     superbly
    0.90
     admirably
    0.90
    better
    0.87
    Activations Density 0.163%

    No Known Activations