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. 14391
    Prev
    Next
    INDEX
    Explanations

    verbs and phrases related to mental processes and contemplation

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "thinking" and its related contexts

    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
    ifer
    -0.77
     Additional
    -0.62
    orer
    -0.61
     Videos
    -0.60
    essee
    -0.60
     Parade
    -0.60
     provide
    -0.58
     Provide
    -0.58
    ector
    -0.58
    ews
    -0.57
    POSITIVE LOGITS
     thinking
    3.56
     Thinking
    2.49
    thinking
    2.36
     imagining
    1.87
     reasoning
    1.78
     wondering
    1.77
     thinkers
    1.63
    think
    1.58
     dreaming
    1.55
     believing
    1.51
    Activations Density 0.019%

    No Known Activations