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. 7851
    Prev
    Next
    INDEX
    Explanations

    sequences of characters related to rare medical terms

    oai_token-act-pair · gpt-3.5-turbo

    references to the concept of “ocean.”

    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
    ĪĴ
    -0.87
     Verse
    -0.75
    ħĭ
    -0.74
     Furious
    -0.66
     lapse
    -0.66
     Chimera
    -0.65
     provocative
    -0.64
     wording
    -0.62
     intention
    -0.61
     Romantic
    -0.61
    POSITIVE LOGITS
    entric
    1.31
    abulary
    1.16
    onut
    1.14
    hester
    1.12
    lean
    1.12
    urrencies
    1.10
    rim
    1.06
    opa
    1.06
    orpor
    1.06
    opy
    1.06
    Activations Density 0.014%

    No Known Activations