© 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. 3-RES-JB
    11. 12488
    Prev
    Next
    INDEX
    Explanations

    references to business entities

    oai_token-act-pair · gpt-3.5-turbo

    references to business entities or companies

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 3-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    3
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    lihood
    -0.78
     Maw
    -0.67
     Died
    -0.67
     Extrem
    -0.66
     Writ
    -0.64
     Strip
    -0.64
     scratch
    -0.64
     Brotherhood
    -0.61
     Viz
    -0.61
     Bagg
    -0.60
    POSITIVE LOGITS
     firms
    1.05
     outsourcing
    0.87
    orpor
    0.84
     firm
    0.79
    arse
    0.76
    folk
    0.74
     specializing
    0.74
    ello
    0.73
     headquartered
    0.73
    cape
    0.72
    Activations Density 0.012%

    No Known Activations