© 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. 4-RES-JB
    11. 20116
    Prev
    Next
    INDEX
    Explanations

    phrases referring to rarity or unusualness

    oai_token-act-pair · gpt-3.5-turbo

    terms that denote rarity or uniqueness, particularly in relation to numerical data and classifications

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

    No Comments

    Negative Logits
    abal
    -0.77
    rums
    -0.75
    lite
    -0.69
    tics
    -0.66
    em
    -0.65
     UNCLASSIFIED
    -0.64
    lich
    -0.63
    omics
    -0.63
    Leaks
    -0.63
    osaurus
    -0.63
    POSITIVE LOGITS
     Seasons
    0.65
     Millenn
    0.59
    romeda
    0.59
    eredith
    0.58
    ¥µ
    0.58
    --+
    0.57
    ?????-
    0.56
    conservancy
    0.56
    fecture
    0.56
    ĵ
    0.56
    Activations Density 0.075%

    No Known Activations