© 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. 3-RES-JB
    11. 12840
    Prev
    Next
    INDEX
    Explanations

    adjectives and nouns related to conflicts or challenges

    oai_token-act-pair · gpt-3.5-turbo

    terms related to social dynamics and organizational structure

    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
    EVA
    -0.58
     respectively
    -0.50
    anwhile
    -0.50
    senal
    -0.50
    uart
    -0.48
    itored
    -0.46
     inaugural
    -0.46
    Thousands
    -0.45
    enegger
    -0.44
     overseen
    -0.43
    POSITIVE LOGITS
    coin
    0.60
     crochet
    0.58
     answ
    0.56
     yourselves
    0.55
    natureconservancy
    0.52
     glim
    0.50
     seiz
    0.49
     cannabin
    0.49
     Morty
    0.49
    gamer
    0.47
    Activations Density 3.473%

    No Known Activations