© 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. 0-RES-JB
    11. 12123
    Prev
    Next
    INDEX
    Explanations

    words related to expressing ideas or viewpoints clearly and coherently

    oai_token-act-pair · gpt-3.5-turbo

    terms related to articulation and clear communication

    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
    piracy
    -0.87
     Rou
    -0.78
     unlucky
    -0.76
    uld
    -0.76
     bonuses
    -0.72
     starve
    -0.71
     server
    -0.70
    worm
    -0.70
     cheating
    -0.69
     Cookies
    -0.66
    POSITIVE LOGITS
     artic
    3.36
     articulate
    3.17
     articulated
    2.75
     eloqu
    1.56
    unci
    1.45
     embodied
    1.37
     embodies
    1.28
     encaps
    1.27
     embod
    1.25
     succinct
    1.24
    Activations Density 0.062%

    No Known Activations