© 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. 3-RES-JB
    11. 23445
    Prev
    Next
    INDEX
    Explanations

    keywords related to gummy candies

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "dummy" and related variations

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 3-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    3
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ledge
    -0.73
    externalActionCode
    -0.72
    auts
    -0.69
    inition
    -0.68
    rawdownloadcloneembedreportprint
    -0.66
    notation
    -0.65
    AME
    -0.65
    lim
    -0.65
    AUT
    -0.64
    leans
    -0.63
    POSITIVE LOGITS
    ummy
    0.88
    ORPG
    0.79
     mummy
    0.72
    bear
    0.71
     Surprise
    0.67
     contag
    0.66
    eus
    0.65
    ipop
    0.65
     Crate
    0.63
    Boss
    0.62
    Activations Density 0.011%

    No Known Activations