© 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. 3-RES-JB
    11. 9218
    Prev
    Next
    INDEX
    Explanations

    references to awards or distinctions

    oai_token-act-pair · gpt-3.5-turbo

    references to military honors and decorations, particularly the Medal of Honor

    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
    minist
    -0.81
    astical
    -0.81
    yg
    -0.80
    ept
    -0.78
    ktop
    -0.78
     skelet
    -0.76
    imposed
    -0.75
    yth
    -0.75
    ynes
    -0.74
    bern
    -0.72
    POSITIVE LOGITS
     Harrington
    0.96
    naire
    0.92
     Qiao
    0.82
    naires
    0.80
     Bravo
    0.75
    Reply
    0.74
     Ventura
    0.70
     Camera
    0.67
     Corps
    0.67
     Canal
    0.67
    Activations Density 0.036%

    No Known Activations