© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-2B
    4. Residual Stream - 16k
    5. 18-GEMMASCOPE-RES-16K
    6. 14203
    Prev
    Next
    INDEX
    Explanations

    phrases related to traditional or nostalgic elements of gaming and technology

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

    Follows a preposition

    np_acts-logits-general · gemini-2.0-flash

    old vs new

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 18-gemmascope-res-16k
    Configuration
    google/gemma-scope-2b-pt-res/layer_18/width_16k/average_l0_74
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.18.hook_resid_post
    Hook Layer
    18
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     CreateTagHelper
    -0.86
    XmlType
    -0.58
    OrBuilder
    -0.55
    zzlement
    -0.53
     moderna
    -0.52
    ullan
    -0.51
    +#+#
    -0.51
    IsContent
    -0.49
     hini
    -0.49
    ばかり
    -0.49
    POSITIVE LOGITS
     traditional
    1.44
    Traditional
    1.31
    traditional
    1.25
     Traditional
    1.24
     physical
    1.13
     traditionnelle
    1.13
     traditionnels
    1.09
     tradicionales
    1.08
     traditionellen
    1.06
     tradicional
    1.06
    Activations Density 0.478%

    No Known Activations