© 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. 4-RES-JB
    11. 20186
    Prev
    Next
    INDEX
    Explanations

    words related to proper nouns or specific entities, potentially related to people, places, or products

    oai_token-act-pair · gpt-3.5-turbo

    suffixes and particular letter combinations in words

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

    No Comments

    Negative Logits
    >)
    -0.56
     ..."
    -0.53
     BIG
    -0.52
     neut
    -0.50
    -0.49
    taboola
    -0.47
     major
    -0.46
    )\
    -0.45
     variance
    -0.45
    amy
    -0.44
    POSITIVE LOGITS
    eers
    0.63
     Canaver
    0.55
    ois
    0.55
    ttes
    0.54
    ese
    0.53
     Chronicles
    0.53
    strap
    0.52
    enance
    0.52
    Redditor
    0.51
    osuke
    0.51
    Activations Density 0.475%

    No Known Activations