© 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. 126704
    Prev
    Next
    INDEX
    Explanations

    analysis

    np_max-act · gemini-2.0-flash

    This neuron detects summary/discourse phrases that introduce findings or conclusions, such as “From this analysis” or “From this study.”

    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
    benchmark
    -0.07
     lumber
    -0.06
    hammer
    -0.06
    .intersection
    -0.06
    _robot
    -0.06
     implant
    -0.06
    �
    -0.06
    ีการ
    -0.06
     Hakk
    -0.06
    陆
    -0.06
    POSITIVE LOGITS
     сна
    0.07
    272
    0.07
     Й
    0.06
     český
    0.06
    ový
    0.06
    ernen
    0.06
     starring
    0.06
     holy
    0.06
    _Character
    0.06
    OPEN
    0.06
    Activations Density 0.031%

    No Known Activations