© 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
    7. Resid Post - 131k
    9. 11-RESID-POST-AA
    11. 126141
    Prev
    Next
    INDEX
    Explanations

    line

    np_max-act · gemini-2.0-flash

    This neuron detects instances of the phrase “in line” (as in “keep … in line”), marking when the text speaks about keeping someone disciplined or under control.

    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
    �
    -0.06
     заболева
    -0.06
     subclass
    -0.06
     subclasses
    -0.06
    ën
    -0.06
    际
    -0.06
     ultra
    -0.06
    ям
    -0.06
    -or
    -0.06
     surplus
    -0.06
    POSITIVE LOGITS
     congregation
    0.07
    播放
    0.07
     Mats
    0.07
    =[]
    0.06
    ainers
    0.06
    ABSPATH
    0.06
    (Resources
    0.06
     RequestContext
    0.06
    (Action
    0.06
    //------------------------------------------------
    0.06
    Activations Density 0.050%

    No Known Activations