© 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. 2-RES-JB
    6. 336
    Prev
    Next
    INDEX
    Explanations

    capitalized proper nouns or acronyms

    oai_token-act-pair · gpt-4-turbo

    proper nouns, particularly from the entity "Quartz", as well as terms related to specific items such as "cider", actions like "crow", and legal terminology like "QC"

    oai_token-act-pair · gpt-3.5-turbo

    references to specific organizations or brands

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 2-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    2
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    igmat
    -0.89
    jah
    -0.79
    heed
    -0.76
    oot
    -0.73
    enges
    -0.73
    gh
    -0.73
    infect
    -0.71
    anger
    -0.71
    reet
    -0.68
    iors
    -0.68
    POSITIVE LOGITS
     cider
    0.93
     Quartz
    0.89
     Flask
    0.85
    perty
    0.78
     crystal
    0.73
    ISM
    0.72
     axes
    0.71
    ADRA
    0.70
     NRS
    0.69
    sonian
    0.69
    Activations Density 0.020%

    No Known Activations