© 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 EvalsExportsinterp-engineNEWAPI 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. 4-RES-JB
    11. 23879
    Prev
    Next
    INDEX
    Explanations

    fantasy races or creatures, such as Orcs, Elves, Dwarves, and Goblins

    oai_token-act-pair · gpt-3.5-turbo

    references to fantasy races and creatures

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

    No Comments

    Negative Logits
    alez
    -0.80
    apeake
    -0.77
    inventoryQuantity
    -0.74
    ettings
    -0.73
    ĥ
    -0.72
    addock
    -0.72
    oulos
    -0.71
    ulate
    -0.69
    ĸļ
    -0.68
    urer
    -0.67
    POSITIVE LOGITS
    ishly
    0.92
     elf
    0.88
     elves
    0.83
    inav
    0.79
    llan
    0.79
     Dwarf
    0.76
    folk
    0.75
     Elf
    0.73
    izabeth
    0.72
    kind
    0.70
    Activations Density 0.023%

    No Known Activations