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. 0-RES-JB
    6. 3710
    Prev
    Next
    INDEX
    Explanations

    references to written content or authorship

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "writes."

    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
     RIS
    -0.64
    gest
    -0.61
    cept
    -0.61
    rium
    -0.60
    erest
    -0.60
     ground
    -0.58
     trailer
    -0.56
     season
    -0.56
     halftime
    -0.55
    frac
    -0.55
    POSITIVE LOGITS
     writes
    3.58
     wrote
    2.29
     write
    1.91
     reads
    1.82
    writ
    1.75
    Writ
    1.66
    wrote
    1.65
     publishes
    1.61
     observes
    1.52
    Writing
    1.49
    Activations Density 0.012%

    No Known Activations