© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 11-RES-JB
    11. 16526
    Prev
    Next
    INDEX
    Explanations

    phrases used for clarification or emphasis, such as "put it simply" or "more succinctly"

    oai_token-act-pair · gpt-3.5-turbo

    phrases that express simplification or clarification of complex ideas

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 11-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre
    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
    incible
    -0.79
    imura
    -0.64
     Skydragon
    -0.63
    ensor
    -0.63
    hung
    -0.62
    anting
    -0.60
    avorite
    -0.60
    awed
    -0.59
     Unable
    -0.59
    lde
    -0.59
    POSITIVE LOGITS
     bluntly
    1.34
     mildly
    1.20
     succinct
    1.19
     plainly
    1.13
     politely
    1.06
     nicely
    1.01
     differently
    0.99
     concise
    0.98
     kindly
    0.94
     diplom
    0.94
    Activations Density 0.078%

    No Known Activations