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. 17499
    Prev
    Next
    INDEX
    Explanations

    the adjective "decent"

    oai_token-act-pair · gpt-3.5-turbo

    the word "decent" and its variations, indicating qualitative assessments

    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
    Enlarge
    -0.72
    ischer
    -0.68
    ï
    -0.66
     Proceedings
    -0.64
    iu
    -0.64
    ius
    -0.63
    ula
    -0.62
     »
    -0.62
    ary
    -0.61
    ELL
    -0.60
    POSITIVE LOGITS
     decent
    3.35
     respectable
    1.93
     reasonable
    1.73
     mediocre
    1.60
     reasonably
    1.47
     good
    1.44
     competent
    1.43
     nice
    1.42
     sane
    1.41
     sensible
    1.38
    Activations Density 0.010%

    No Known Activations