© 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. 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. 121211
    Prev
    Next
    INDEX
    Explanations

    satire and parodies

    np_max-act · gemini-2.0-flash

    tokens that mark chat structure and role/metadata (system/user/assistant markers, start/end boundaries and other formatting/quote/punctuation markers).

    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
    一步
    -0.07
    万名
    -0.06
    Prev
    -0.06
    有的
    -0.06
    prev
    -0.06
    对自己的
    -0.06
     мен
    -0.06
    Days
    -0.06
    .parent
    -0.06
    rypted
    -0.06
    POSITIVE LOGITS
     Vacuum
    0.08
    信仰
    0.07
     boutique
    0.07
     ToString
    0.07
     hardly
    0.07
    classified
    0.07
    ɧ
    0.07
     ghost
    0.06
     myśli
    0.06
    eled
    0.06
    Activations Density 0.125%

    No Known Activations