© 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. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    5. GPT2-Small
    7. Residual Stream
    9. 0-RES-JB
    11. 17996
    Prev
    Next
    INDEX
    Explanations

    the word "Magic" accompanied by a numerical value in the context of various subjects, such as games, music, novels, and community activities

    oai_token-act-pair · gpt-3.5-turbo

    mentions of "Magic" in various contexts, likely related to the "Magic: The Gathering" game or associated media

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 0-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.hook_resid_pre
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    Data Type
    torch.float32
    Hook Point
    blocks.0.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    0
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    Assad
    -0.98
    houn
    -0.79
    onde
    -0.74
     Assad
    -0.73
    ansk
    -0.72
    anchester
    -0.72
    ndum
    -0.72
    ilitary
    -0.70
     nep
    -0.70
     diver
    -0.68
    POSITIVE LOGITS
     Magic
    3.79
    Magic
    3.18
    magic
    2.33
     magic
    2.05
     Wizards
    2.02
     Magical
    1.76
     Mages
    1.57
     magician
    1.56
     Wizard
    1.48
     Mage
    1.48
    Activations Density 0.006%

    No Known Activations