© 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. 11-RES-JB
    11. 9879
    Prev
    Next
    INDEX
    Explanations

    instances where someone is being criticized for making a mistake or an error in judgment

    oai_token-act-pair · gpt-3.5-turbo

    common themes of mistakes or misconceptions in understanding or approaches

    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
    idable
    -0.72
    orem
    -0.71
    ivably
    -0.70
    ãĥ´ãĤ¡
    -0.68
     Pry
    -0.67
    venge
    -0.66
     Sever
    -0.64
     Heights
    -0.64
    bolt
    -0.64
    shall
    -0.63
    POSITIVE LOGITS
     simplistic
    1.28
     misconceptions
    1.17
     misunderstanding
    1.11
     precon
    1.08
     privile
    1.08
     stereotypes
    1.08
     reliance
    1.06
     stereotyp
    1.05
     assumptions
    1.04
     misinformation
    1.04
    Activations Density 0.705%

    No Known Activations