© 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. 11-RES-JB
    11. 1875
    Prev
    Next
    INDEX
    Explanations

    phrases related to someone speaking

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "says."

    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
    cffffcc
    -0.77
    ampion
    -0.72
    illus
    -0.70
    aired
    -0.65
    à¦
    -0.64
    rendered
    -0.64
    isoft
    -0.64
    gallery
    -0.63
    icut
    -0.63
     pione
    -0.62
    POSITIVE LOGITS
     he
    0.85
     goodbye
    0.80
     thats
    0.79
     that
    0.77
     there
    0.77
     they
    0.76
     it
    0.72
     she
    0.68
     nobody
    0.67
     hello
    0.65
    Activations Density 0.156%

    No Known Activations