© 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. 1-RES-JB
    11. 13136
    Prev
    Next
    INDEX
    Explanations

    names related to a character called "Charlie Mage."

    oai_token-act-pair · gpt-3.5-turbo

    references to the concept of "mage" and related characters or entities

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 1-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    1
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ometown
    -0.83
    ervatives
    -0.82
    anke
    -0.81
    irds
    -0.77
    oho
    -0.76
    avez
    -0.75
    rities
    -0.75
    ervative
    -0.74
    osher
    -0.74
    biased
    -0.73
    POSITIVE LOGITS
    llan
    1.43
     Mage
    1.04
     mage
    0.94
    sonian
    0.90
     mages
    0.88
    mons
    0.81
    aign
    0.76
    ept
    0.76
    craft
    0.75
    maid
    0.73
    Activations Density 0.012%

    No Known Activations