© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Qwen3.5-4B
    3. 15-RES-MATRYOSHKA-65K
    4. 18949
    Prev
    Next
    INDEX
    Explanations

    This is happening because the necessary input lists (`MAX_ACTIVATING_TOKENS`, `TOKENS_AFTER_MAX_ACTIVATING_TOKEN`, `TOP_POSITIVE_LOGITS`, `TOP_ACTIVATING_TEXTS`) were empty or incomplete in the prompt. To provide an explanation, I need some data from these lists to identify patterns.Please provide the content for the lists so I can analyze the neuron's behavior

    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
    oga
    -0.07
    way
    -0.06
    lyn
    -0.06
    ëį°
    -0.05
    olla
    -0.05
     approach
    -0.05
    zyn
    -0.05
    奴婢
    -0.05
    iders
    -0.05
    ivel
    -0.05
    POSITIVE LOGITS
    orado
    0.07
    æ¯ĴçĹĩ
    0.06
    essant
    0.06
    å¿Ĺ强
    0.06
    èŃī
    0.06
     Affid
    0.06
    ç¾İ
    0.05
    ensee
    0.05
     stat
    0.05
    -browser
    0.05
    Activations Density 0.038%

    No Known Activations