© 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. 8-RES-JB
    11. 2019
    Prev
    Next
    INDEX
    Explanations

    phrases related to conflict and resolution

    oai_token-act-pair · gpt-3.5-turbo

    confrontational or combative language

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Hass
    -0.66
     Magnus
    -0.62
    thodox
    -0.61
    utical
    -0.60
    isconsin
    -0.59
    ©¶æ
    -0.59
    ricane
    -0.56
    igs
    -0.55
    agascar
    -0.54
    annis
    -0.53
    POSITIVE LOGITS
     automatically
    0.88
     immediately
    0.85
     corresponding
    0.77
     instantly
    0.73
     promptly
    0.68
     subconscious
    0.67
     magically
    0.66
     invol
    0.65
    mediately
    0.64
     accordingly
    0.64
    Activations Density 1.141%

    No Known Activations