© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    2. Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
    3. Llama3.1-8B-IT (Instruct)
    4. Resid Post - 131k
    5. 11-RESID-POST-AA
    6. 41548
    Prev
    Next
    INDEX
    Explanations

    conversational text

    np_max-act · gemini-2.0-flash

    This neuron detects high-level control or “jailbreak” instructions—i.e. explicit user commands that override prior rules or ask the model to adopt a new mode.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    andyrdt/saes-llama-3.1-8b-instruct/resid_post_layer_11/trainer_1
    Dataset (Dashboard)
    Various
    Features
    131,072
    Data Type
    float32
    Hook Name
    blocks.11.hook_resid_post
    Architecture
    standard
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     ventured
    -0.07
     linewidth
    -0.07
     înt
    -0.06
    	aux
    -0.06
    674
    -0.06
     توس
    -0.06
    Thomas
    -0.06
     roots
    -0.06
     Hash
    -0.06
    unicorn
    -0.06
    POSITIVE LOGITS
     cái
    0.07
    toThrow
    0.07
     specified
    0.07
    Writable
    0.07
    _RPC
    0.06
    hamster
    0.06
    кое
    0.06
    кому
    0.06
    ASH
    0.06
     sağlam
    0.06
    Activations Density 0.008%

    No Known Activations