© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW 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. 80326
    Prev
    Next
    INDEX
    Explanations

    's

    np_max-act · gemini-2.0-flash

    contractions where a pronoun or "there" is followed by an apostrophe and a clitic (I'm, he's, she's, we're, there's, it's, etc.).

    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
    集
    -0.07
     lucky
    -0.07
     Mitar
    -0.07
     insured
    -0.07
     ///</
    -0.07
    之一
    -0.06
    |
    -0.06
    얻
    -0.06
    幸福
    -0.06
    atable
    -0.06
    POSITIVE LOGITS
     coeffs
    0.07
     зуб
    0.07
    ܘ
    0.07
    Глав
    0.07
    ]+'
    0.07
    𝗽
    0.07
    gross
    0.07
    _ENSURE
    0.07
    赃
    0.07
     langue
    0.07
    Activations Density 0.429%

    No Known Activations