© 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
    3. Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
    5. Llama3.1-8B-IT (Instruct)
    7. Resid Post - 131k
    9. 11-RESID-POST-AA
    11. 109859
    Prev
    Next
    INDEX
    Explanations

    expressions and sayings

    np_max-act · gemini-2.0-flash

    The neuron activates specifically on tokens from the assistant’s reply segments (distinguishing assistant output from user or system text).

    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
    мещ
    -0.07
    _regs
    -0.07
     blinking
    -0.07
     believed
    -0.07
    vä
    -0.06
    .LEFT
    -0.06
    /watch
    -0.06
     renewable
    -0.06
    Dependency
    -0.06
    Dream
    -0.06
    POSITIVE LOGITS
     exploits
    0.08
    .model
    0.07
     Louisville
    0.06
     si
    0.06
     whichever
    0.06
     произ
    0.06
     '))↵
    0.06
    .full
    0.06
    /trans
    0.06
    )","
    0.06
    Activations Density 0.064%

    No Known Activations