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. 21123
    Prev
    Next
    INDEX
    Explanations

    common verbs that denote the act of concentrating or paying special attention to a specific thing or concept

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "focus" in various 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
    ares
    -0.76
    hazard
    -0.72
    76561
    -0.66
     illegally
    -0.64
    illes
    -0.63
    tre
    -0.63
    SU
    -0.62
    DN
    -0.61
     treacher
    -0.61
    inside
    -0.61
    POSITIVE LOGITS
     focus
    3.68
    focus
    2.59
     emphasis
    2.23
     focusing
    2.11
     concentrate
    2.10
     Focus
    2.09
    Focus
    2.08
     focused
    2.00
     focuses
    1.97
     foc
    1.80
    Activations Density 0.013%

    No Known Activations