© 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. 0-RES-JB
    11. 7642
    Prev
    Next
    INDEX
    Explanations

    references to the name "Michael" with variations in spelling and context

    oai_token-act-pair · gpt-3.5-turbo

    the word "hel" and its variations, which likely relate to specific names or terms in the context of the document

    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
     chant
    -0.70
     swear
    -0.65
     indisc
    -0.63
     inflation
    -0.63
     vintage
    -0.62
     duplicate
    -0.61
     print
    -0.60
     sophistic
    -0.59
     Diplom
    -0.58
     SN
    -0.58
    POSITIVE LOGITS
    hel
    4.68
    helial
    2.16
    Hel
    1.66
    her
    1.60
    hell
    1.60
    helm
    1.55
    helle
    1.49
    hem
    1.47
    hest
    1.46
    het
    1.43
    Activations Density 0.021%

    No Known Activations