© 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. 13432
    Prev
    Next
    INDEX
    Explanations

    retrospect

    np_max-act · gemini-2.0-flash

    The neuron detects words and phrases that signal looking back or reflecting on past events (e.g., “in retrospect,” “hindsight,” “retrospectively”).

    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
     Covenant
    -0.07
    _MSG
    -0.07
     Electro
    -0.07
     Venture
    -0.07
    DIC
    -0.07
    Minimum
    -0.06
     Assy
    -0.06
    ฎ
    -0.06
    ooth
    -0.06
     Gaussian
    -0.06
    POSITIVE LOGITS
     hindsight
    0.08
     retrospect
    0.08
     BCM
    0.07
    $link
    0.07
     ViewBag
    0.06
    анной
    0.06
    author
    0.06
     Hon
    0.06
     xsi
    0.06
     subscribed
    0.06
    Activations Density 0.005%

    No Known Activations