© 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. 0-RES-JB
    11. 10198
    Prev
    Next
    INDEX
    Explanations

    specific numerical values

    oai_token-act-pair · gpt-3.5-turbo

    numerical values related to specific IDs or codes

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 0-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    0
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     âĶľ
    -0.84
     Revel
    -0.81
     repl
    -0.79
     Pug
    -0.78
     Pyrrha
    -0.75
     Vermont
    -0.74
     Weaver
    -0.74
     Riv
    -0.73
     Benson
    -0.73
    ãĤ©
    -0.73
    POSITIVE LOGITS
    214
    2.56
    213
    2.02
    215
    2.00
     215
    1.80
     214
    1.79
    212
    1.76
    216
    1.72
    211
    1.64
    217
    1.63
     213
    1.59
    Activations Density 0.034%

    No Known Activations