© 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. 10329
    Prev
    Next
    INDEX
    Explanations

    phrases related to ending a sentence or a thought

    oai_token-act-pair · gpt-3.5-turbo

    sentences that end with a period

    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
     ambush
    -0.66
     manif
    -0.66
     niece
    -0.64
     granddaughter
    -0.63
     nuts
    -0.62
     grandson
    -0.61
     tracing
    -0.60
     lamb
    -0.59
     mate
    -0.59
     stranger
    -0.59
    POSITIVE LOGITS
    tumblr
    0.99
    google
    0.97
     Please
    0.90
    assetsadobe
    0.90
    aspx
    0.90
    jpg
    0.85
    php
    0.84
    txt
    0.84
    taboola
    0.83
     Cannot
    0.82
    Activations Density 0.208%

    No Known Activations