Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
    3. Qwen2.5-7B-IT
    4. Resid Post - 131k
    5. 15-RESID-POST-AA
    6. 115312
    Prev
    Next
    INDEX
    Explanations

    Dialogue/conversations

    np_max-act · gemini-2.0-flash

    Tokens that begin an assistant reply or dialogue turn—especially opening/confirmation phrases like "Of", "Of course," and the leading quote marks that start a response.

    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
     tele
    -0.07
    .Package
    -0.07
    Pokemon
    -0.07
     dword
    -0.06
     Staples
    -0.06
     nak
    -0.06
    .market
    -0.06
    公然
    -0.06
     Identified
    -0.06
    🎼
    -0.06
    POSITIVE LOGITS
     aux
    0.08
    밞
    0.07
    衝
    0.07
     האמיתי
    0.07
    โรค
    0.07
    ела
    0.07
    ᨵ
    0.07
    овых
    0.07
    ภาษา
    0.07
    夫妇
    0.07
    Activations Density 0.117%

    No Known Activations