© 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. 19640
    Prev
    Next
    INDEX
    Explanations

    numerical values ending with the number '120'

    oai_token-act-pair · gpt-3.5-turbo

    numerical values, particularly related to measurements or quantities

    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
    friend
    -0.85
    gart
    -0.82
    folk
    -0.80
    giving
    -0.77
    tip
    -0.75
    let
    -0.74
    ling
    -0.72
    gary
    -0.70
    doms
    -0.70
    ster
    -0.69
    POSITIVE LOGITS
    803
    0.87
    cffff
    0.84
     ILCS
    0.83
    mph
    0.80
    508
    0.80
    dB
    0.79
    456
    0.78
    807
    0.77
    505
    0.76
    595
    0.75
    Activations Density 0.018%

    No Known Activations