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. 84739
    Prev
    Next
    INDEX
    Explanations

    commands/code/programming

    np_max-act · gemini-2.0-flash

    This neuron responds to tokens that mark bot commands and their syntax—such as command names, exclamation-prefixes, angle-bracket parameters, and related command‐format words.

    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
    ئ
    -0.07
    棒
    -0.07
    ildo
    -0.07
    _ptrs
    -0.07
    ्वत
    -0.06
     fins
    -0.06
    君
    -0.06
     Tam
    -0.06
     ylim
    -0.06
     dp
    -0.06
    POSITIVE LOGITS
     searcher
    0.08
    executable
    0.06
    改革
    0.06
     fascinating
    0.06
    arsers
    0.06
    actice
    0.06
     Christianity
    0.06
     jadx
    0.06
     στι
    0.06
     securely
    0.06
    Activations Density 0.019%

    No Known Activations