© 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. 0-RES-JB
    11. 19005
    Prev
    Next
    INDEX
    Explanations

    specific words in a pattern about watching repeated content

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "Replay" indicating video content

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 0-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    0
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     sulph
    -0.78
     opium
    -0.65
     rounded
    -0.63
     Imperial
    -0.63
     adjunct
    -0.62
     stout
    -0.62
    âĢIJ
    -0.62
    §
    -0.61
     bod
    -0.61
     employed
    -0.60
    POSITIVE LOGITS
     Replay
    4.58
     replay
    1.88
     WATCHED
    1.28
    JUST
    1.05
    `,
    1.04
     repl
    1.01
    Reviewer
    0.93
     Refresh
    0.89
     Echoes
    0.88
     Rew
    0.87
    Activations Density 0.014%

    No Known Activations