© 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. 2087
    Prev
    Next
    INDEX
    Explanations

    terms related to political actions or discussions

    oai_token-act-pair · gpt-3.5-turbo

    the word 'basically' and related concepts involving basic elements or principles

    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
     Adren
    -0.70
    \/\/
    -0.69
     vict
    -0.69
     Roberts
    -0.68
     stars
    -0.67
     Cannes
    -0.64
     Corona
    -0.63
     attractions
    -0.62
     clerks
    -0.61
     oy
    -0.61
    POSITIVE LOGITS
    flow
    1.87
     curfew
    1.52
    bas
    1.47
    ovan
    1.35
     ceasefire
    1.14
    reddit
    1.13
    utan
    1.12
    verb
    1.02
     veiled
    0.95
    olving
    0.95
    Activations Density 0.027%

    No Known Activations