© 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. 12-RES-JB
    11. 4705
    Prev
    Next
    INDEX
    Explanations

    references to historical events or trends

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot

    phrases referring to time periods or durations

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 12-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_post
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    Data Type
    torch.float32
    Hook Point
    blocks.11.hook_resid_post
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    11
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     pione
    -0.95
    ą
    -0.87
     RandomRedditor
    -0.87
    ė
    -0.87
    ö
    -0.87
    Ě
    -0.87
    oreAndOnline
    -0.87
    Č
    -0.87
    ġ
    -0.87
    Á
    -0.87
    POSITIVE LOGITS
     decade
    1.49
     few
    1.39
     couple
    1.35
     year
    1.21
     fortnight
    1.19
     several
    1.18
     month
    1.16
     dozen
    1.15
     century
    1.12
     week
    1.11
    Activations Density 0.056%

    No Known Activations