© 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. 0-RES-JB
    11. 15249
    Prev
    Next
    INDEX
    Explanations

    terms related to molecular biology or molecular-level processes

    oai_token-act-pair · gpt-3.5-turbo

    terms related to molecular processes and structures

    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
    FB
    -0.71
    Js
    -0.67
    orse
    -0.66
     âľ
    -0.66
    hett
    -0.64
    ikk
    -0.63
    reddit
    -0.63
    Seg
    -0.62
    bucks
    -0.62
     Songs
    -0.61
    POSITIVE LOGITS
     molecular
    3.35
     Molecular
    2.25
     biochemical
    1.48
     genomic
    1.32
     biomedical
    1.31
     cellular
    1.18
     microsc
    1.17
     computational
    1.17
     microscopic
    1.17
     biological
    1.16
    Activations Density 0.015%

    No Known Activations