© 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. 1-RES-JB
    11. 22942
    Prev
    Next
    INDEX
    Explanations

    the word "shouldn't."

    oai_token-act-pair · gpt-3.5-turbo

    the word "shouldn't" and its variations, indicating a focus on negations or prohibitions

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 1-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    1
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Chains
    -0.71
    hunt
    -0.68
    ULT
    -0.67
    lip
    -0.65
     Puzzle
    -0.64
    band
    -0.64
    icted
    -0.61
    activated
    -0.60
    dress
    -0.59
     Mutant
    -0.59
    POSITIVE LOGITS
    't
    1.18
    \'
    0.92
     rightfully
    0.89
    ['
    0.82
    FontSize
    0.81
     applause
    0.73
     shouldn
    0.72
     rightly
    0.72
    (*
    0.71
    DragonMagazine
    0.69
    Activations Density 0.005%

    No Known Activations