© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 30710
    Prev
    Next
    INDEX
    Explanations

    conversational text excerpts

    np_max-act · gemini-2.0-flash

    The neuron strongly activates on tokens that are part of simple inline numeric expressions—especially the slash “/” in fractions (e.g. “1/” in “1/1” or similarly formatted numeric snippets).

    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
     forwarded
    -0.06
     зав
    -0.06
    orption
    -0.06
    ConfigurationException
    -0.06
     recommends
    -0.06
    ifferential
    -0.06
     preceding
    -0.06
     종
    -0.06
    }↵↵↵↵↵↵
    -0.06
     pas
    -0.06
    POSITIVE LOGITS
    	map
    0.07
     cata
    0.06
     zajím
    0.06
     srdce
    0.06
    _positions
    0.06
    CAST
    0.06
    ΙΑΣ
    0.06
    にと
    0.06
    něte
    0.06
     unins
    0.06
    Activations Density 0.132%

    No Known Activations