© 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 (Instruct)
    7. Resid Post - 131k
    9. 11-RESID-POST-AA
    11. 110148
    Prev
    Next
    INDEX
    Explanations

    mathematical notation

    np_max-act · gemini-2.0-flash

    The neuron activates on LaTeX math‐mode tokens (e.g. “\langle”, “rangle”, subscripts, Greek letters) especially in expectation‐value notation.

    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
     Lav
    -0.07
     logic
    -0.07
    ﻟ
    -0.06
     культури
    -0.06
    totalCount
    -0.06
     bilg
    -0.06
    _FAULT
    -0.06
    发现
    -0.06
    ترنت
    -0.06
    ۰۰۰
    -0.06
    POSITIVE LOGITS
     enforcing
    0.07
    ña
    0.07
    .accept
    0.07
     singular
    0.07
     indebted
    0.06
     brain
    0.06
     anom
    0.06
    ’an
    0.06
     paralysis
    0.06
     astounding
    0.06
    Activations Density 0.002%

    No Known Activations