© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    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. 3-RES-JB
    11. 24096
    Prev
    Next
    INDEX
    Explanations

    instances of the word "analyze"

    oai_token-act-pair · gpt-3.5-turbo

    occurrences of the word "analyze" and its variants

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 3-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    3
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    birth
    -0.79
    bearing
    -0.72
    guard
    -0.69
    fired
    -0.69
    hood
    -0.69
    hold
    -0.66
    Lago
    -0.65
    vic
    -0.64
    guards
    -0.63
    \\\\
    -0.62
    POSITIVE LOGITS
     analy
    0.94
     analyze
    0.85
     analyzing
    0.83
     analys
    0.83
    racuse
    0.81
    eering
    0.79
     analyse
    0.78
    alyses
    0.78
    analy
    0.76
    rador
    0.75
    Activations Density 0.023%

    No Known Activations