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. 104162
    Prev
    Next
    INDEX
    Explanations

    Correctness/negation

    np_max-act · gemini-2.0-flash

    Neuron 4 activates on warning or caution language—especially the “if not handled correctly” pattern—i.e. it flags mentions of improper handling safety warnings.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8

    instructions and safety guidelines for handling and using chemicals in the industry.

    oai_token-act-pair · gpt-4o-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
     peppers
    -0.08
    yclopedia
    -0.07
     chips
    -0.07
    KANJI
    -0.06
     EDM
    -0.06
     Everybody
    -0.06
     hamstring
    -0.06
    _PRINTF
    -0.06
    702
    -0.06
     Priest
    -0.06
    POSITIVE LOGITS
     backdrop
    0.07
    应用
    0.07
     сохран
    0.06
     نخست
    0.06
     dx
    0.06
     interv
    0.06
    -points
    0.06
    ็จ
    0.06
    decor
    0.06
    stock
    0.06
    Activations Density 0.013%

    No Known Activations