© 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
    2. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    3. GPT2-Small
    4. Residual Stream
    5. 0-RES-JB
    6. 15776
    Prev
    Next
    INDEX
    Explanations

    phrases related to the name "Sabrina" and the term "Katrina."

    oai_token-act-pair · gpt-3.5-turbo

    occurrences of specific names or references, particularly related to a person named Sabrina and mentions of Hurricane Katrina

    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
     curves
    -0.69
     FIG
    -0.69
     curve
    -0.66
     bodies
    -0.65
     blocks
    -0.63
     boards
    -0.63
     plaque
    -0.63
     steroids
    -0.63
     Ack
    -0.63
     Patent
    -0.60
    POSITIVE LOGITS
    rina
    4.66
     Katrina
    1.23
    obyl
    1.07
    hens
    1.04
    onica
    0.99
    untu
    0.99
    erella
    0.99
     Isa
    0.97
    acci
    0.93
    riz
    0.93
    Activations Density 0.051%

    No Known Activations