Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit 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. 15898
    Prev
    Next
    INDEX
    Explanations

    specific texts that contain the word "Consensus" with varying levels of intensity (9 or 10) depending on the context

    oai_token-act-pair · gpt-3.5-turbo

    the term "cons" and related usages, likely indicative of consumer-related content or discussions

    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
     cutoff
    -0.73
     pounding
    -0.72
     teasing
    -0.70
     stub
    -0.68
     tunnel
    -0.67
     glare
    -0.65
     doorway
    -0.65
    baugh
    -0.64
     bullied
    -0.64
     squee
    -0.64
    POSITIVE LOGITS
    Cons
    3.85
    cons
    2.41
     Cons
    2.24
     CONS
    1.66
     cons
    1.48
    Pros
    1.42
    Con
    1.38
     Consumers
    1.24
    CON
    1.24
    Consumer
    1.19
    Activations Density 0.007%

    No Known Activations