Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    3. GPT2-Small
    4. Residual Stream
    5. 2-RES-JB
    6. 9559
    Prev
    Next
    INDEX
    Explanations

    The token "Re" at the beginning of a word, often followed by other text, frequently appearing after an end-of-text marker. This suggests it might be part of a metadata field or a system-generated tag, possibly related to replies, responses, or re-posts.

    eleuther_acts_top20 · gemini-1.5-flashTriggered by @johnny

    specific snippets formatted as "Re" followed by a number

    oai_token-act-pair · gpt-3.5-turbo

    occurrences of the substring "Re" in various contexts

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 2-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    2
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Nicarag
    -0.71
     CPR
    -0.70
     à¨
    -0.68
     Contra
    -0.67
    ãĥ¼ãĥĨãĤ£
    -0.66
     Anchorage
    -0.66
     CPC
    -0.66
     Butterfly
    -0.65
     Icelandic
    -0.64
    åŃ
    -0.64
    POSITIVE LOGITS
    becca
    1.26
    ceived
    1.12
    verend
    1.06
    venge
    1.04
    levant
    1.03
    uben
    1.01
    asons
    1.01
    agan
    1.00
    ggie
    1.00
    asonable
    0.98
    Activations Density 0.010%

    No Known Activations