© 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. 124734
    Prev
    Next
    INDEX
    Explanations

    Web content and questions

    np_max-act · gemini-2.0-flash

    user question turns in a chat transcript, i.e., message boundaries and interrogative prompts indicating a new user query.

    oai_token-act-pair · gpt-5Triggered by @vetterc0

    markers of chat turn boundaries and inter-message metadata (non-linguistic separator lines), especially right after a user message ends.

    oai_token-act-pair · gpt-5Triggered 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
    Urban
    -0.07
    俱
    -0.07
    畸形
    -0.06
     Mặt
    -0.06
    数字化
    -0.06
    同业
    -0.06
    UnityEngine
    -0.06
    }:${
    -0.06
    (Class
    -0.06
     oste
    -0.06
    POSITIVE LOGITS
    imen
    0.08
    KS
    0.07
    imming
    0.07
    ımı
    0.07
    contrib
    0.07
     tearing
    0.07
    /delete
    0.07
     Liam
    0.07
    berries
    0.07
     çözüm
    0.07
    Activations Density 0.091%

    No Known Activations