© 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 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. 8-RES-JB
    11. 19281
    Prev
    Next
    INDEX
    Explanations

    references to direct speech with strong emotional expressions

    oai_token-act-pair · gpt-3.5-turbo

    instances of the article "a" and various uses of it in the context of conversation and language

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    evidence
    -0.79
    Attach
    -0.73
    Edit
    -0.71
    Events
    -0.70
     Ankara
    -0.70
    Interested
    -0.70
    Autom
    -0.69
    Ask
    -0.69
    Attempts
    -0.68
    Contents
    -0.68
    POSITIVE LOGITS
     lot
    1.41
     bunch
    1.41
     guy
    1.08
     shitty
    1.07
     nice
    1.04
     little
    1.02
     crappy
    1.02
     thousand
    1.01
     whole
    1.01
     damn
    1.00
    Activations Density 0.646%

    No Known Activations