Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    3. GPT2-Small
    4. Residual Stream
    5. 1-RES-JB
    6. 1691
    Prev
    Next
    INDEX
    Explanations

    instances where the word "Down" is used in a context that likely refers to the genetic disorder Down syndrome

    oai_token-act-pair · gpt-3.5-turbo

    mentions of "Down" in various contexts, particularly relating to Down syndrome and individuals associated with it

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 1-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    1
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ĸļ
    -0.79
    aceous
    -0.70
    ature
    -0.69
    ================================
    -0.68
    andum
    -0.68
    ifling
    -0.66
    olkien
    -0.65
    archives
    -0.64
     fabricated
    -0.63
     dear
    -0.62
    POSITIVE LOGITS
    stairs
    1.54
    graded
    0.94
    hill
    0.92
    stem
    0.91
     Down
    0.88
    grades
    0.88
    Down
    0.87
    LOAD
    0.86
    pour
    0.85
    lap
    0.84
    Activations Density 0.007%

    No Known Activations