Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit 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. 8-RES-JB
    6. 5678
    Prev
    Next
    INDEX
    Explanations

    words related to emotional responses and actions, such as disappointment, excitement, delight, shock, disgust, pleasure, and blaming

    oai_token-act-pair · gpt-3.5-turbo

    prefixes or stems related to negative emotions or reactions

    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
     sheets
    -0.67
     Beetle
    -0.65
     Solitaire
    -0.65
     Puzzles
    -0.63
     ðŁĻĤ
    -0.62
     Masquerade
    -0.61
     follower
    -0.61
     bake
    -0.60
     targ
    -0.60
     Odin
    -0.60
    POSITIVE LOGITS
    ificantly
    1.20
    actly
    1.17
    initely
    1.14
    mediately
    1.09
    etheless
    1.07
    appropriately
    1.06
    ighed
    1.03
    viously
    1.01
    ufficient
    0.99
    redibly
    0.99
    Activations Density 0.203%

    No Known Activations