© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    5. GPT2-Small
    7. Residual Stream
    9. 0-RES-JB
    11. 4851
    Prev
    Next
    INDEX
    Explanations

    phrases indicating harm or negative impact

    oai_token-act-pair · gpt-3.5-turbo

    words related to grievances or negative feelings

    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
     staples
    -0.65
     mustard
    -0.63
     Stretch
    -0.62
    ync
    -0.62
     clutter
    -0.61
     catchy
    -0.60
     clinch
    -0.60
     dope
    -0.59
     promise
    -0.59
     lockdown
    -0.59
    POSITIVE LOGITS
    rieved
    3.54
    cession
    1.45
     harmed
    1.40
    vernment
    1.33
    rogram
    1.00
    spective
    0.98
     offended
    0.96
    verning
    0.94
    ivalent
    0.94
    imilar
    0.94
    Activations Density 0.047%

    No Known Activations