© 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. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    5. GPT2-Small
    7. Residual Stream
    9. 11-RES-JB
    11. 23781
    Prev
    Next
    INDEX
    Explanations

    words related to magic tricks or objects associated with magic.

    oai_token-act-pair · gpt-3.5-turbo

    references to the concept of magic

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 11-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    11
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    dated
    -0.68
    VIEW
    -0.67
     Fas
    -0.64
    ppers
    -0.64
    lishes
    -0.61
    aches
    -0.61
    rontal
    -0.60
    views
    -0.60
     FEC
    -0.59
    isco
    -0.57
    POSITIVE LOGITS
     wand
    0.93
     Leap
    0.90
     carpet
    0.88
     Beans
    0.87
     mushrooms
    0.86
    ãĥĥãĤ¯
    0.86
    ian
    0.83
     realism
    0.82
     tricks
    0.80
     beans
    0.78
    Activations Density 0.048%

    No Known Activations