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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 10-RES-JB
    11. 12792
    Prev
    Next
    INDEX
    Explanations

    information or details related specifically to technology and search queries

    oai_token-act-pair · gpt-3.5-turbo

    references to searching for or accessing information

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 10-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    10
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     alas
    -0.67
     onwards
    -0.65
     reluctantly
    -0.64
     onward
    -0.63
     panicked
    -0.63
     hardly
    -0.62
     bitterly
    -0.62
     moreover
    -0.62
    hemoth
    -0.61
    Thankfully
    -0.61
    POSITIVE LOGITS
     subreddits
    0.83
    oplan
    0.79
     topics
    0.76
    natureconservancy
    0.75
     pokemon
    0.73
     SCP
    0.72
     Featured
    0.71
     keywords
    0.70
    ratom
    0.69
     KDE
    0.67
    Activations Density 0.663%

    No Known Activations