© 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. 6318
    Prev
    Next
    INDEX
    Explanations

    references to different types of berries

    oai_token-act-pair · gpt-3.5-turbo

    instances of the suffix "ries" and related variations, indicating a focus on the word pattern involving plural nouns, particularly in contexts related to categories or items

    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
    CVE
    -0.74
    yrinth
    -0.72
    earable
    -0.69
    agonists
    -0.67
     informants
    -0.66
    ilogy
    -0.64
    awar
    -0.64
     meanings
    -0.63
     inval
    -0.62
     stewards
    -0.62
    POSITIVE LOGITS
    ries
    1.04
    lander
    0.92
    gaard
    0.87
    EF
    0.84
    ried
    0.82
    stal
    0.79
    xtap
    0.78
    vere
    0.77
    cks
    0.76
    ymes
    0.74
    Activations Density 0.010%

    No Known Activations