© 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
    2. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    3. GPT2-Small
    4. Residual Stream
    5. 0-RES-JB
    6. 13946
    Prev
    Next
    INDEX
    Explanations

    government websites

    oai_token-act-pair · gpt-3.5-turbo

    references to government websites or entities

    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
     Moonlight
    -0.75
     tail
    -0.68
     skins
    -0.68
     scal
    -0.67
     gra
    -0.66
     shaving
    -0.65
     belts
    -0.64
     Phant
    -0.64
     translation
    -0.63
     Sword
    -0.63
    POSITIVE LOGITS
    gov
    4.66
    Gov
    1.92
    edu
    1.62
    org
    1.38
    government
    1.27
    nih
    1.25
    com
    1.06
    Government
    1.05
     Citiz
    1.02
    biz
    0.98
    Activations Density 0.012%

    No Known Activations