© 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. 4-RES-JB
    11. 4619
    Prev
    Next
    INDEX
    Explanations

    proper nouns related to a specific individual named "Hart" and possibly variations of the name like "hardt"

    oai_token-act-pair · gpt-3.5-turbo

    mentions of people, particularly those with the suffix 'hart' or 'hardt'

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

    No Comments

    Negative Logits
    URI
    -0.78
    ntil
    -0.76
     millenn
    -0.73
     veter
    -0.71
     sclerosis
    -0.70
     tampering
    -0.67
    Loading
    -0.65
    pty
    -0.64
     sarc
    -0.64
     Palestin
    -0.64
    POSITIVE LOGITS
    hart
    1.05
    enstein
    0.97
    ford
    0.95
    ãĤ¤ãĥĪ
    0.87
    awi
    0.83
    mann
    0.81
    ness
    0.81
    hardt
    0.81
    bone
    0.81
    ilton
    0.80
    Activations Density 0.015%

    No Known Activations