© 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. 70916
    Prev
    Next
    INDEX
    Explanations

    The neuron mainly detects first-person/self-referential language — tokens that signal the speaker's personal statements or opinions (e.g., "I", "I'm", "we", "really").

    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
     Bios
    -0.07
    照料
    -0.07
     dữ
    -0.07
    tır
    -0.06
     Decide
    -0.06
    夔
    -0.06
     txn
    -0.06
    知情
    -0.06
    Oper
    -0.06
    茨
    -0.06
    POSITIVE LOGITS
     ex
    0.08
    ♙
    0.07
     bağlı
    0.07
    (change
    0.07
     Celtics
    0.07
    잖아요
    0.07
    (found
    0.06
    planet
    0.06
    endars
    0.06
    意大利
    0.06
    Activations Density 0.226%

    No Known Activations