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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 1-RES-JB
    11. 9414
    Prev
    Next
    INDEX
    Explanations

    verbal commands or instructions starting with "don't."

    oai_token-act-pair · gpt-3.5-turbo

    negations or expressions of inability

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 1-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    1
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     behavi
    -0.80
     tremend
    -0.77
    EStream
    -0.76
     mosqu
    -0.75
     ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³
    -0.73
     gorilla
    -0.69
     intern
    -0.69
     cannabin
    -0.68
     exha
    -0.66
     Skydragon
    -0.66
    POSITIVE LOGITS
    ween
    1.07
    aken
    1.06
    otally
    1.04
    asks
    1.04
    ruck
    1.03
    ractor
    1.02
    olkien
    1.01
    akers
    1.00
    ople
    0.97
    une
    0.97
    Activations Density 0.127%

    No Known Activations