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. 1-RES-JB
    6. 8115
    Prev
    Next
    INDEX
    Explanations

    words related to waste management, particularly focusing on the word "trash"

    oai_token-act-pair · gpt-3.5-turbo

    references to waste and debris

    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
    akening
    -0.92
     Archdemon
    -0.80
    arine
    -0.71
    thren
    -0.66
    intent
    -0.65
     organising
    -0.64
    iator
    -0.63
    ths
    -0.63
     operating
    -0.63
    slow
    -0.63
    POSITIVE LOGITS
     trash
    1.08
     Trash
    0.98
     dumps
    0.98
    dump
    0.94
     cans
    0.91
     heap
    0.90
    bags
    0.88
    ãģį
    0.88
     bags
    0.87
    bag
    0.86
    Activations Density 0.007%

    No Known Activations