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. Llama3.1-8B-IT (Instruct)
    4. Resid Post - 131k
    5. 11-RESID-POST-AA
    6. 55666
    Prev
    Next
    INDEX
    Explanations

    Pressure, science

    np_max-act · gemini-2.0-flash

    This neuron detects the assistant’s speaker‐label marker (the “<|start_header_id|>” token indicating an assistant response).

    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
    oked
    -0.07
    _TAB
    -0.07
    _pose
    -0.07
    persona
    -0.07
     setters
    -0.06
     negotiate
    -0.06
    utsche
    -0.06
    hash
    -0.06
     acad
    -0.06
     Plum
    -0.06
    POSITIVE LOGITS
     prostřednictvím
    0.07
    므로
    0.07
     enlarge
    0.06
     Circular
    0.06
    .Alert
    0.06
    Thank
    0.06
     Factors
    0.06
    .innerHTML
    0.06
     použití
    0.06
     sposób
    0.06
    Activations Density 0.032%

    No Known Activations