© 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. 0-RES-JB
    11. 15119
    Prev
    Next
    INDEX
    Explanations

    phrases related to negative outcomes or failures

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "losing."

    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
     Inform
    -0.70
    æĹ
    -0.67
    Instruct
    -0.66
    chrome
    -0.64
    fu
    -0.61
     Embassy
    -0.61
    include
    -0.60
     INFO
    -0.60
    Offic
    -0.59
    rote
    -0.59
    POSITIVE LOGITS
     losing
    3.45
     lose
    1.96
     winning
    1.84
     loss
    1.75
    loss
    1.73
     loses
    1.72
     losses
    1.71
     losers
    1.58
     lost
    1.57
     gaining
    1.53
    Activations Density 0.020%

    No Known Activations