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

    Neuronpedia

    APICircuit 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. 4-RES-JB
    6. 8146
    Prev
    Next
    INDEX
    Explanations

    adjectives related to negative or disastrous events

    oai_token-act-pair · gpt-3.5-turbo

    terms related to catastrophic events and disasters

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

    No Comments

    Negative Logits
    elf
    -0.76
    yard
    -0.75
    cius
    -0.75
    pel
    -0.75
    quart
    -0.74
    hung
    -0.74
    ulton
    -0.71
    appa
    -0.71
    bors
    -0.71
    stra
    -0.70
    POSITIVE LOGITS
     consequences
    0.98
    itous
    0.96
     havoc
    0.95
     catastrophic
    0.88
     impacts
    0.84
     consequence
    0.82
     outcomes
    0.82
     catast
    0.81
     endanger
    0.81
     nightmares
    0.81
    Activations Density 0.016%

    No Known Activations