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. 11-RES-JB
    6. 15930
    Prev
    Next
    INDEX
    Explanations

    words related to close and positive relationships or bonds between entities

    oai_token-act-pair · gpt-3.5-turbo

    phrases related to interpersonal relationships and connections

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 11-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    11
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    cens
    -0.78
    warning
    -0.73
     Amnesty
    -0.71
    regon
    -0.70
    ocaust
    -0.69
     Shutterstock
    -0.67
    terday
    -0.67
     Demand
    -0.67
     Disabled
    -0.66
    veland
    -0.65
    POSITIVE LOGITS
     friendship
    1.92
     friendships
    1.87
     relationship
    1.63
     relationships
    1.60
    relations
    1.58
     acquaintances
    1.49
     mutual
    1.47
     intimacy
    1.47
     rapport
    1.45
     partnership
    1.43
    Activations Density 0.695%

    No Known Activations