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

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Qwen3.5-4B
    5. 15-RES-MATRYOSHKA-65K
    7. 20854
    Prev
    Next
    INDEX
    Explanations

    **Analysis of Neuron Behavior**The neuron appears to be sensitive to the start of quotations or sentences, often followed by words indicating a new event, statement, or subject.**Explanation:**start of quotes or sentences

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    decoderesearch/qwen-3.5-saes/qwen-3.5-4b
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     sometime
    -0.06
    ÅĽciÄħ
    -0.05
    æŁł
    -0.05
    çµ®
    -0.05
    yt
    -0.05
    ç»°
    -0.05
    缴æݥ被
    -0.05
    è¿ĩæĿ¥
    -0.05
    æħİ
    -0.05
    IDL
    -0.05
    POSITIVE LOGITS
    ache
    0.06
    isku
    0.05
    å½¢æĢģ
    0.05
     ach
    0.05
    åħ¥éĢī
    0.05
     vari
    0.05
    èĢģåıĭ
    0.05
    itus
    0.05
    äufer
    0.05
    çĤ¹èµŀ
    0.05
    Activations Density 0.063%

    No Known Activations