Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & 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. 5355
    Prev
    Next
    INDEX
    Explanations

    terms related to technical processes and systems involving cooling and heating mechanisms.

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

    information related to technical processes and systems, particularly in the fields of refrigeration, medical devices, and editing suites.

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot

    physical or mechanical components of systems or structures, particularly focusing on functional parts like chambers, motors, rooms, or containers.

    oai_token-act-pair · claude-3-5-sonnet-20240620Triggered by @bot

    preceding punctuation or conjunctions

    np_acts-logits-general · gemini-2.0-flash

    via or followed by conjunctions

    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
     FetchType
    -0.76
     terhadap
    -0.68
    égard
    -0.66
     into
    -0.63
     unto
    -0.62
    CppMethod
    -0.59
     compared
    -0.58
    providedIn
    -0.57
     estekak
    -0.56
    Fprintf
    -0.56
    POSITIVE LOGITS
     via
    0.99
     tramite
    0.74
     vía
    0.74
    via
    0.73
     terdekat
    0.72
    为止
    0.67
    去了
    0.64
    Via
    0.64
    يكب
    0.63
     VIA
    0.62
    Activations Density 0.823%

    No Known Activations