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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 7923
    Prev
    Next
    INDEX
    Explanations

    words related to death or dying

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "dies" and related expressions of death or failure

    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
    itton
    -0.64
    uca
    -0.64
    aver
    -0.61
    ophy
    -0.59
    ounty
    -0.58
     simmer
    -0.58
    ranging
    -0.58
    oured
    -0.58
     Taste
    -0.57
    vy
    -0.56
    POSITIVE LOGITS
     dies
    3.08
     fails
    2.09
     ceases
    1.67
     loses
    1.51
     survives
    1.51
     suffers
    1.49
     collapses
    1.48
     succeeds
    1.44
     kills
    1.38
     misses
    1.37
    Activations Density 0.023%

    No Known Activations