© 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. 2207
    Prev
    Next
    INDEX
    Explanations

    phrases indicating a request or demand to take a specific action

    oai_token-act-pair · gpt-3.5-turbo

    requests or demands expressed with the word "to."

    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
     intended
    -0.62
    ãĤ§
    -0.62
    forcing
    -0.60
     Laksh
    -0.59
     caused
    -0.59
     plots
    -0.59
     Rider
    -0.58
     meant
    -0.58
     tipped
    -0.57
    rophe
    -0.57
    POSITIVE LOGITS
     differ
    1.14
     borrow
    1.08
     speak
    1.03
     know
    1.02
     see
    0.98
     hear
    0.96
     postpone
    0.91
    pless
    0.88
     remain
    0.88
     meet
    0.87
    Activations Density 0.117%

    No Known Activations