© 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 EvalsExportsinterp-engineNEWAPI 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. 11-RES-JB
    11. 11805
    Prev
    Next
    INDEX
    Explanations

    the word "don't" or "don't" contraction

    oai_token-act-pair · gpt-3.5-turbo

    phrases expressing negation or refusal

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 11-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    11
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Osc
    -0.75
     Maritime
    -0.73
     Dickinson
    -0.69
     Liberty
    -0.67
     Compass
    -0.64
     Lisbon
    -0.64
     Spectrum
    -0.61
     Robinson
    -0.60
     Quart
    -0.57
     Tripoli
    -0.56
    POSITIVE LOGITS
    âĢ
    1.19
    Í
    1.11
     condone
    0.97
     forget
    0.90
     disappoint
    0.90
     apologise
    0.90
     discriminate
    0.89
     apologize
    0.89
     âĢ
    0.87
     bother
    0.87
    Activations Density 0.132%

    No Known Activations