© 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. 8-RES-JB
    11. 6988
    Prev
    Next
    INDEX
    Explanations

    words related to written correspondence such as letters, emails, and memos

    oai_token-act-pair · gpt-3.5-turbo

    references to written communications such as letters, memos, and statements

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     races
    -0.65
     encounters
    -0.61
     journeys
    -0.61
    iencies
    -0.59
     corridors
    -0.59
     Yards
    -0.58
    ittens
    -0.58
     breeds
    -0.57
     extremes
    -0.57
     Cups
    -0.56
    POSITIVE LOGITS
     titled
    1.03
     stating
    0.95
     thanking
    0.93
     entitled
    0.90
     saying
    0.90
     outlining
    0.89
     apologizing
    0.87
     urging
    0.86
     wherein
    0.85
     denouncing
    0.84
    Activations Density 0.228%

    No Known Activations