© 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
    3. Gemma-3-12B
    5. 24-GEMMASCOPE-2-RES-16K
    7. 12517
    Prev
    Next
    INDEX
    Explanations

    S-box, hydration, ahead, rogue, executed, ECU, allow

    np_acts-logits-general · gemini-2.5-flash-lite

    The neuron fires most strongly on uncommon, low‐frequency words (i.e. rare vocabulary items).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    extremely long, technical, or encyclopedic text passages that are atypical of natural conversational writing.

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

    prepositions that link phrases, especially those signaling purpose or accompaniment.

    oai_token-act-pair · gpt-5Triggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-pt/resid_post/layer_24_width_16k_l0_medium
    Prompts (Dashboard)
    392,802 prompts, 256 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    rs
    0.72
    can
    0.69
     can
    0.66
    utional
    0.66
    m
    0.65
    ዎችን
    0.64
    oretically
    0.63
    ️
    0.63
    es
    0.62
     Mountains
    0.62
    POSITIVE LOGITS
    ﻖ
    0.96
    하기
    0.86
    ΠΑ
    0.81
    ﻢ
    0.80
    𝑮
    0.80
    으로
    0.78
    한다
    0.78
    чай
    0.77
    हाउस
    0.77
    하여
    0.77
    Activations Density 0.000%

    No Known Activations