© 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. 14659
    Prev
    Next
    INDEX
    Explanations

    words indicating a source or method of obtaining information

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "via" used to indicate sources or methods of information

    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
     stand
    -0.71
     leve
    -0.70
    urance
    -0.64
     ord
    -0.64
     tightened
    -0.63
     drawn
    -0.63
     row
    -0.63
     hal
    -0.62
     fault
    -0.61
     wrench
    -0.61
    POSITIVE LOGITS
    via
    4.51
    Via
    2.21
     Via
    1.36
     via
    1.22
    thanks
    1.20
    advertising
    1.13
    through
    1.06
    isu
    1.02
    fax
    0.99
    vs
    0.96
    Activations Density 0.010%

    No Known Activations