© 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. Llama3.1-8B-IT
    7. Resid Post - 131k
    9. 15-RESID-POST-AA
    11. 26117
    Prev
    Next
    INDEX
    Explanations

    Q\&A forum snippets

    np_max-act · gemini-2.0-flash

    This neuron detects special conversation metadata and speaker/header tokens (the model's control tokens that mark message boundaries and roles).

    oai_token-act-pair · gpt-5-miniTriggered by @vetterc0
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    andyrdt/saes-llama-3.1-8b-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
    龍
    -0.07
    ерг
    -0.07
     '</
    -0.06
    러스
    -0.06
     picked
    -0.06
    ヶ月
    -0.06
    넷
    -0.06
    seg
    -0.06
    Dallas
    -0.06
     чет
    -0.06
    POSITIVE LOGITS
     './../../
    0.07
     AssetImage
    0.06
     russe
    0.06
    illance
    0.06
    \Entity
    0.06
    _weights
    0.06
    ailable
    0.06
     اصول
    0.06
    leshooting
    0.06
    -----------
    0.06
    Activations Density 0.047%

    No Known Activations