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. 3-RES-JB
    6. 8440
    Prev
    Next
    INDEX
    Explanations

    phrases related to advanced or leading-edge technology or practices

    oai_token-act-pair · gpt-3.5-turbo

    references to "cutting edge" concepts or advancements

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 3-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    3
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ña
    -0.79
     Prosecut
    -0.69
    ech
    -0.69
    cia
    -0.68
    sama
    -0.66
    hma
    -0.65
    Registered
    -0.65
     Monitor
    -0.65
    eon
    -0.64
    igmat
    -0.64
    POSITIVE LOGITS
    lass
    1.10
    scenes
    1.09
     torches
    1.01
    scene
    1.00
    lasses
    0.99
    cliffe
    0.86
    thro
    0.85
     edge
    0.83
    cutting
    0.83
     cutting
    0.81
    Activations Density 0.016%

    No Known Activations