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. 19622
    Prev
    Next
    INDEX
    Explanations

    words related to the concept of self-harm or self-endangerment

    oai_token-act-pair · gpt-3.5-turbo

    terms related to the concept of "antics" or "antagonistic behavior."

    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
     Roh
    -0.72
     ______
    -0.69
     shake
    -0.66
     kil
    -0.66
     center
    -0.64
     Haz
    -0.63
     glove
    -0.62
     hum
    -0.61
     chancellor
    -0.61
     hid
    -0.61
    POSITIVE LOGITS
    antic
    4.75
    antically
    2.18
    antics
    2.12
    ANT
    1.75
    antis
    1.74
    ant
    1.54
    antine
    1.41
    anta
    1.38
    ancy
    1.36
    antes
    1.25
    Activations Density 0.009%

    No Known Activations