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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 10-RES-JB
    11. 8167
    Prev
    Next
    INDEX
    Explanations

    phrases related to criminal activities, law enforcement, and physical altercations

    oai_token-act-pair · gpt-3.5-turbo

    sequences related to criminal activity and police actions

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 10-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    10
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     publicly
    -0.67
     internationally
    -0.66
    udget
    -0.66
     outsourcing
    -0.66
    Week
    -0.66
     Franchise
    -0.66
     Historically
    -0.63
     Represent
    -0.63
     Institutes
    -0.63
     debates
    -0.63
    POSITIVE LOGITS
     braking
    1.23
     landing
    1.07
     manoeuv
    1.06
     collision
    1.05
     collided
    1.00
     graz
    1.00
     rapp
    0.99
     parach
    0.98
     debris
    0.98
     immobil
    0.96
    Activations Density 0.633%

    No Known Activations