Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2025
    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. 125453
    Prev
    Next
    INDEX
    Explanations

    sincerity or truthfulness

    np_max-act · gemini-2.0-flash

    This neuron responds to words and phrases that convey embarrassment, hesitation, or reluctance (e.g., “embarrassed,” “didn’t,” “odd that he…”).

    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
    .te
    -0.07
    _ANS
    -0.06
     insecurity
    -0.06
     darken
    -0.06
    town
    -0.06
    (TABLE
    -0.06
    	if
    -0.06
     bracelets
    -0.06
    	main
    -0.06
     draft
    -0.06
    POSITIVE LOGITS
     glGen
    0.07
    èles
    0.07
     EVAL
    0.07
    ้ม
    0.07
     decl
    0.07
     bv
    0.07
    χρι
    0.06
    ैज
    0.06
     Dön
    0.06
    〉
    0.06
    Activations Density 0.084%

    No Known Activations