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. 3711
    Prev
    Next
    INDEX
    Explanations

    words related to different types of shifts, such as work shifts, schedule shifts, and phase shifts

    oai_token-act-pair · gpt-3.5-turbo

    references to work shifts and their variations

    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
     treasure
    -0.73
     guarantee
    -0.63
     sale
    -0.63
    ument
    -0.63
     gift
    -0.63
    ston
    -0.61
     bundle
    -0.61
    example
    -0.61
     blessed
    -0.61
    www
    -0.60
    POSITIVE LOGITS
     shifts
    3.64
     swings
    1.72
     transitions
    1.65
     changes
    1.52
     shifting
    1.47
     shift
    1.45
     surges
    1.44
     declines
    1.43
     flips
    1.39
     switches
    1.39
    Activations Density 0.009%

    No Known Activations