© 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. 22585
    Prev
    Next
    INDEX
    Explanations

    words that have been suffixed with '-ed' or '-ing'

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "of" and related phrases or constructs in a text

    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
    ilyn
    -0.82
     Dise
    -0.78
    ãĤ¤ãĥĪ
    -0.76
    esm
    -0.75
    iage
    -0.69
    ngth
    -0.67
    artifacts
    -0.65
     Pradesh
    -0.64
    awi
    -0.63
    hess
    -0.63
    POSITIVE LOGITS
     aloud
    0.90
     poem
    0.79
     encoding
    0.74
     abbre
    0.74
     verse
    0.73
     poems
    0.73
     phrases
    0.73
     describing
    0.71
     acronym
    0.69
     poetic
    0.69
    Activations Density 0.337%

    No Known Activations