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

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. OpenMOSS · Llama Scope: SAEs for Llama-3.1-8B
    5. Llama3.1-8B (Base)
    7. Residual Stream
    9. 15-LLAMASCOPE-RES-131K
    11. 88085
    Prev
    Next
    INDEX
    Explanations

    references to sharing content and expressing reluctance to reveal sensitive information

    oai_token-act-pair · gpt-4o-miniTriggered by @bot

    Phrases like "don't want to," "would," or other modal verbs paired with explanations of why certain content is being withheld, typically to maintain appropriateness of the text or to avoid spoilers.

    eleuther_acts_top20 · claude-3-7-sonnet-20250219Triggered by @lu-christina
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    cerebras/SlimPajama-627B
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     bulk
    -0.06
    ij
    -0.06
    ainless
    -0.05
    ilee
    -0.05
     Bulk
    -0.05
    asks
    -0.05
     experimental
    -0.05
    aina
    -0.05
    BK
    -0.05
    iso
    -0.05
    POSITIVE LOGITS
    ÙĤÛĮ
    0.07
    ERE
    0.07
    SENS
    0.07
    ëĭī
    0.07
     Kaf
    0.07
    imin
    0.07
    ruk
    0.07
    UCE
    0.06
    dess
    0.06
    TRACE
    0.06
    Activations Density 0.015%

    No Known Activations