© 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. 2-RES-JB
    11. 23949
    Prev
    Next
    INDEX
    Explanations

    phrases related to specific character names, especially those including the word "Elf."

    oai_token-act-pair · gpt-3.5-turbo

    references to elves and related character names

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 2-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    2
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    Īè
    -0.76
    MSN
    -0.69
     Chapters
    -0.66
    ļéĨĴ
    -0.64
     Chrom
    -0.64
    âĸ¬
    -0.63
     Tile
    -0.63
    Rail
    -0.62
     Ou
    -0.62
     Purchase
    -0.62
    POSITIVE LOGITS
    efe
    1.07
    inition
    1.00
    bre
    0.84
    lesh
    0.83
    iture
    0.83
    cffff
    0.82
    redit
    0.82
    gypt
    0.82
    lich
    0.79
    yre
    0.78
    Activations Density 0.016%

    No Known Activations