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. 26544
    Prev
    Next
    INDEX
    Explanations

    reason

    np_max-act · gemini-2.0-flash

    phrases asking for reasons or explanations.

    oai_token-act-pair · gpt-4oTriggered by @tcai

    This neuron detects question cues asking for reasons or justifications (e.g. “reason,” “why,” “any good reason,” “recommended way”).

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

    discussions about sexual attraction and the complexities surrounding it.

    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
    .authentication
    -0.07
    シュ
    -0.06
    -New
    -0.06
     căn
    -0.06
    enes
    -0.06
    τεύ
    -0.06
     Signs
    -0.06
    —a
    -0.06
     ing
    -0.06
     BEST
    -0.06
    POSITIVE LOGITS
    /source
    0.06
     tersebut
    0.06
    ]"↵
    0.06
     :↵↵↵↵
    0.06
    .faces
    0.06
     düş
    0.06
    ?.
    0.06
     ----------------
    0.06
    lift
    0.06
    .compare
    0.06
    Activations Density 0.077%

    No Known Activations