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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & 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. 81455
    Prev
    Next
    INDEX
    Explanations

    thoughts

    np_max-act · gemini-2.0-flash

    The neuron activates on words that refer to sharing or soliciting input—like “thoughts,” “ideas,” “feedback,” and “insights.”

    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
     army
    -0.06
    	reload
    -0.06
     animation
    -0.06
     illusion
    -0.06
    Interfaces
    -0.06
     arbitration
    -0.06
     punish
    -0.06
     جو
    -0.06
    'a
    -0.06
     гру
    -0.06
    POSITIVE LOGITS
     thoughts
    0.09
     Odds
    0.07
     fantast
    0.07
     tq
    0.06
    íc
    0.06
    PS
    0.06
    maxcdn
    0.06
     창
    0.06
     viewpoints
    0.06
    식을
    0.06
    Activations Density 0.013%

    No Known Activations