© 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
    3. Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
    5. Qwen2.5-7B-IT
    7. Resid Post - 131k
    9. 15-RESID-POST-AA
    11. 12212
    Prev
    Next
    INDEX
    Explanations

    Lake

    np_max-act · gemini-2.0-flash

    This neuron detects explicit taboo content—sexual/violent and bodily-excretion terms (especially incestuous or otherwise extreme/offensive sexual phrasing).

    oai_token-act-pair · gpt-5-miniTriggered by @vetterc0
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    andyrdt/saes-qwen2.5-7b-instruct/resid_post_layer_15/trainer_1
    Dataset (Dashboard)
    Various
    Features
    131,072
    Data Type
    float32
    Hook Name
    blocks.15.hook_resid_post
    Architecture
    standard
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ultiply
    -0.07
    分别为
    -0.07
     phosph
    -0.07
     provincia
    -0.07
    他の
    -0.06
     [].
    -0.06
    不仅可以
    -0.06
    畜牧
    -0.06
     enlightened
    -0.06
    使用权
    -0.06
    POSITIVE LOGITS
    pliance
    0.08
     reliable
    0.07
     binaries
    0.07
    baby
    0.07
     payroll
    0.07
     tun
    0.07
    مس
    0.07
     Pad
    0.06
     lookup
    0.06
     replay
    0.06
    Activations Density 0.196%

    No Known Activations