© 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. 11-RES-JB
    11. 21807
    Prev
    Next
    INDEX
    Explanations

    influence and impact-related words with a strong connotation of magnitude

    oai_token-act-pair · gpt-3.5-turbo

    words indicating large amounts or significant impacts

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 11-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    11
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ilk
    -0.77
    ©¶æ
    -0.77
    »Ĵ
    -0.75
    roma
    -0.74
    athon
    -0.71
    yssey
    -0.70
    athetic
    -0.69
     effortlessly
    -0.68
    handle
    -0.67
    fest
    -0.66
    POSITIVE LOGITS
     amounts
    1.26
     quantities
    1.05
     difficulties
    1.05
     disparities
    1.03
     parallels
    1.00
     challenges
    1.00
     ramifications
    0.99
     discrepancies
    0.98
     repercussions
    0.98
     delays
    0.98
    Activations Density 0.156%

    No Known Activations