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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-9B
    4. Residual Stream - 16k
    5. 31-GEMMASCOPE-RES-16K
    6. 6655
    Prev
    Next
    INDEX
    Explanations

    legal terminology and concepts related to judicial proceedings

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

    modal verbs and auxiliary verbs expressing possibility, necessity, or conditionality (such as "could," "should," "need," "have").

    oai_token-act-pair · claude-4-5-haikuTriggered by @emiglarou

    modal verbs or auxiliary verbs in conditional, hypothetical, or contrastive contexts (like "could," "should," "must," "did" when expressing possibility, obligation, or contrary-to-fact situations).

    oai_token-act-pair · claude-4-5-sonnetTriggered by @jyhe0408

    auxiliary and modal verb constructions expressing obligation, possibility, or assertion, often in formal/legal prose and negated phrases.

    oai_token-act-pair · gpt-5Triggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B @ 31-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-pt-res/layer_31/width_16k/average_l0_114
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.31.hook_resid_post
    Hook Layer
    31
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    同志
    -0.34
     égard
    -0.31
     herum
    -0.30
     rodillas
    -0.30
     honra
    -0.30
     devant
    -0.30
    GOTREF
    -0.29
     ervoor
    -0.28
     decimos
    -0.28
     alegría
    -0.27
    POSITIVE LOGITS
     neither
    0.90
     nowhere
    0.81
     nothing
    0.81
     NOTHING
    0.75
     nada
    0.67
    neither
    0.65
     Neither
    0.65
     none
    0.64
     Nothing
    0.63
    NOTHING
    0.61
    Activations Density 0.435%

    No Known Activations