© 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. 2812
    Prev
    Next
    INDEX
    Explanations

    phrases indicating a contrast or contradiction compared to a previous expectation

    oai_token-act-pair · gpt-3.5-turbo

    expressions emphasizing contradiction or opposing viewpoints

    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
    bers
    -0.73
    utters
    -0.67
    artney
    -0.65
    asp
    -0.65
    haw
    -0.65
    don
    -0.65
    mberg
    -0.60
    bs
    -0.60
    edu
    -0.59
     Atlas
    -0.59
    POSITIVE LOGITS
     contrary
    1.10
    etheless
    1.03
     undermin
    0.97
     guiActiveUn
    0.84
     notwithstanding
    0.81
    rary
    0.79
     materially
    0.79
    lihood
    0.78
    theless
    0.77
     thereto
    0.76
    Activations Density 0.007%

    No Known Activations