© 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. 2-RES-JB
    11. 17536
    Prev
    Next
    INDEX
    Explanations

    numerical information related to nutrition facts

    oai_token-act-pair · gpt-3.5-turbo

    quantitative measurements related to health or nutrition

    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
     Corpus
    -0.80
    gotten
    -0.76
     Horizons
    -0.69
    Redditor
    -0.68
    ESCO
    -0.66
     jan
    -0.66
    FORE
    -0.66
     Collider
    -0.63
     Galileo
    -0.63
     Victim
    -0.62
    POSITIVE LOGITS
    mt
    1.12
    mg
    1.01
    omez
    0.94
    irez
    0.91
    emonic
    0.89
    allery
    0.89
    fen
    0.88
    ient
    0.87
    ilit
    0.86
    ertodd
    0.86
    Activations Density 0.012%

    No Known Activations