© 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. 4-RES-JB
    11. 9768
    Prev
    Next
    INDEX
    Explanations

    language indicating a need to comprehend or explaining a concept

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "understand."

    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
    onies
    -0.71
    unal
    -0.66
     bounced
    -0.64
    nar
    -0.64
    ammy
    -0.63
    cake
    -0.62
     presided
    -0.62
    onto
    -0.62
     Rath
    -0.61
     stunt
    -0.61
    POSITIVE LOGITS
     WHY
    1.03
     Understanding
    0.93
     how
    0.90
     why
    0.88
     Understand
    0.85
    ably
    0.82
    ledge
    0.80
     understanding
    0.76
     HOW
    0.75
     understand
    0.75
    Activations Density 0.030%

    No Known Activations