© 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. 10-RES-JB
    11. 8440
    Prev
    Next
    INDEX
    Explanations

    dialogue interruptions or interjections

    oai_token-act-pair · gpt-3.5-turbo
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 10-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    10
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     spons
    -0.85
     diving
    -0.82
     exporting
    -0.82
     streng
    -0.81
     forging
    -0.80
     redes
    -0.79
     inheritance
    -0.79
    etheless
    -0.78
     enriched
    -0.78
     centrally
    -0.78
    POSITIVE LOGITS
    Narr
    1.54
    Scene
    1.42
    Cut
    1.37
    Interview
    1.34
    Answer
    1.33
    Pause
    1.33
    Dialogue
    1.28
    Guest
    1.28
    Suddenly
    1.27
    Everyone
    1.25
    Activations Density 0.130%

    No Known Activations