© 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. 7-RES-JB
    11. 2232
    Prev
    Next
    INDEX
    Explanations

    proper nouns or terms related to political and social issues, especially discussing repeals, bills, and public debates

    oai_token-act-pair · gpt-3.5-turbo

    words related to the concept of repetition or revealing information

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 7-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    7
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ãĥĥãĤ¯
    -0.71
     nerv
    -0.68
    ucky
    -0.64
     Manufacturer
    -0.63
    ohyd
    -0.62
     floor
    -0.62
     dust
    -0.60
     wrists
    -0.59
     Haram
    -0.59
    reddits
    -0.59
    POSITIVE LOGITS
    ety
    1.06
    erness
    0.99
    llers
    0.93
    aling
    0.90
    edy
    0.86
    als
    0.86
    esy
    0.84
    ivably
    0.83
    aler
    0.83
    ller
    0.83
    Activations Density 0.067%

    No Known Activations