© 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. 8-RES-JB
    11. 17237
    Prev
    Next
    INDEX
    Explanations

    key phrases or terms used in discourse or dialogue, often indicating a continuation or transition in speech

    oai_token-act-pair · gpt-3.5-turbo

    conversational expressions and interjections related to discourse markers

    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
     airs
    -0.67
    çīĪ
    -0.67
     seiz
    -0.66
     Coffee
    -0.64
     HQ
    -0.63
     Beg
    -0.62
     Obj
    -0.62
     Extension
    -0.61
     Views
    -0.59
     Amb
    -0.59
    POSITIVE LOGITS
    because
    1.21
    unless
    1.18
    but
    1.13
    that
    1.06
    especially
    1.06
    although
    1.05
    whether
    1.04
    where
    1.03
    which
    1.03
    since
    1.02
    Activations Density 0.246%

    No Known Activations