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. 4-RES-JB
    6. 17560
    Prev
    Next
    INDEX
    Explanations

    words related to contrasting statements or additional information

    oai_token-act-pair · gpt-3.5-turbo

    conjunctions, transitions, and qualifiers that suggest contrast or complexity in statements

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

    No Comments

    Negative Logits
     NOW
    -0.43
    gress
    -0.43
    arde
    -0.42
    NBA
    -0.42
    rier
    -0.41
    geist
    -0.40
    DERR
    -0.40
     ASAP
    -0.40
    Tomorrow
    -0.39
     someday
    -0.39
    POSITIVE LOGITS
    pires
    0.49
    kered
    0.48
    ricular
    0.45
     accompanies
    0.44
    ciplinary
    0.42
    ipment
    0.40
    ilateral
    0.39
    pired
    0.39
     huh
    0.38
     occurs
    0.38
    Activations Density 2.390%

    No Known Activations