© 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. 19235
    Prev
    Next
    INDEX
    Explanations

    words related to downgrading or devaluing something

    oai_token-act-pair · gpt-3.5-turbo

    references to credit ratings and their downgrades

    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
     ut
    -0.71
     mask
    -0.66
     trademark
    -0.66
     thunder
    -0.66
    ĸļ
    -0.65
     handle
    -0.63
     articulate
    -0.62
     voice
    -0.62
     understatement
    -0.59
     stir
    -0.59
    POSITIVE LOGITS
    graded
    3.79
    grading
    3.11
    grades
    2.97
    dating
    2.76
    dates
    2.62
    dated
    2.24
    gradation
    1.77
    grade
    1.30
    haul
    1.28
    stairs
    1.25
    Activations Density 0.021%

    No Known Activations