© 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. 10016
    Prev
    Next
    INDEX
    Explanations

    words related to individuals and their relationships or actions

    oai_token-act-pair · gpt-3.5-turbo

    possessive pronouns and references to individuals involved in personal relationships

    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
    liga
    -0.84
    Games
    -0.80
    ãĥ´ãĤ¡
    -0.80
    deals
    -0.77
    trop
    -0.76
    flex
    -0.75
    adan
    -0.74
    hib
    -0.74
    Beast
    -0.74
     [|
    -0.74
    POSITIVE LOGITS
     whereabouts
    1.34
     belongings
    1.34
     assailant
    1.23
     assailants
    1.23
     relatives
    1.14
     own
    1.13
     cousin
    1.13
     alleged
    1.11
     estranged
    1.11
     attacker
    1.10
    Activations Density 0.327%

    No Known Activations