© 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. 18163
    Prev
    Next
    INDEX
    Explanations

    phrases related to excessive or overwhelming amounts

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "Too."

    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
     commence
    -0.66
     inform
    -0.65
     prospect
    -0.64
     composite
    -0.64
     services
    -0.62
     conv
    -0.60
     correspond
    -0.60
     visiting
    -0.59
     visitation
    -0.59
     corresponding
    -0.58
    POSITIVE LOGITS
     Too
    3.41
    Too
    3.17
    too
    2.30
     Enough
    1.53
     too
    1.49
     Wrong
    1.41
    Enough
    1.35
     Worse
    1.27
     Already
    1.13
    Wr
    1.13
    Activations Density 0.017%

    No Known Activations