© 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
    2. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    3. GPT2-Small
    4. Residual Stream
    5. 4-RES-JB
    6. 772
    Prev
    Next
    INDEX
    Explanations

    email subject lines and details about people from different sources

    oai_token-act-pair · gpt-3.5-turbo

    subjects or topics indicated in written correspondence

    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
     relegation
    -0.73
    ctors
    -0.70
     Blaz
    -0.69
    adiator
    -0.66
     suff
    -0.66
    alach
    -0.65
     Gravity
    -0.65
     Falling
    -0.65
    @#&
    -0.65
    itans
    -0.64
    POSITIVE LOGITS
     memos
    1.26
     briefed
    1.24
     briefings
    1.21
    CLASSIFIED
    1.18
     aides
    1.15
     Emails
    1.12
     memo
    1.10
     briefing
    1.07
     aide
    1.05
     confid
    1.03
    Activations Density 0.695%

    No Known Activations