© 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. 3-RES-JB
    11. 16283
    Prev
    Next
    INDEX
    Explanations

    the concept of an idea or hypothesis being put forward or assumed as a basis for discussion or argument

    oai_token-act-pair · gpt-3.5-turbo

    concepts related to foundational ideas or arguments being presented

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 3-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    3
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    vor
    -0.86
    sung
    -0.82
    vals
    -0.72
    eder
    -0.71
    alm
    -0.68
    ocker
    -0.67
    dos
    -0.67
    abbling
    -0.66
     Journals
    -0.66
    ipop
    -0.64
    POSITIVE LOGITS
     premise
    1.30
    REC
    0.80
     underpin
    0.74
     proposition
    0.71
     assumption
    0.70
    terson
    0.68
    ually
    0.67
    lessly
    0.67
    IRE
    0.66
    itial
    0.65
    Activations Density 0.009%

    No Known Activations