© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    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. 12904
    Prev
    Next
    INDEX
    Explanations

    references to moderators, particularly in the context of debates or discussions

    oai_token-act-pair · gpt-3.5-turbo

    references to moderators in discussions or events

    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
     passports
    -0.78
     Yards
    -0.73
     transfers
    -0.70
     Pass
    -0.69
    ranch
    -0.67
     Vert
    -0.66
    utions
    -0.66
     Eff
    -0.65
     Jump
    -0.64
    VEN
    -0.63
    POSITIVE LOGITS
     moderator
    3.96
     moderators
    3.41
     moder
    2.59
     Moder
    1.87
    icia
    1.15
     subreddit
    1.12
     interviewer
    1.11
     facilit
    1.09
    Redditor
    1.07
     moderation
    1.06
    Activations Density 0.030%

    No Known Activations