© 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. 0-RES-JB
    11. 2697
    Prev
    Next
    INDEX
    Explanations

    phrases related to characteristics or enhancements in gaming or other competitive contexts

    oai_token-act-pair · gpt-3.5-turbo

    references to "buff" or improvements in various contexts

    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
     Sears
    -0.81
    ISA
    -0.78
    latest
    -0.77
    violent
    -0.74
    lihood
    -0.70
    operated
    -0.70
     Investigations
    -0.69
    Thom
    -0.67
     Cannabis
    -0.67
    ILS
    -0.67
    POSITIVE LOGITS
     buff
    3.94
     buffs
    2.94
    buff
    2.66
     Buff
    2.52
    Buff
    2.07
     ner
    1.73
     buffers
    1.25
     buffer
    1.18
     deb
    1.05
    Buffer
    1.04
    Activations Density 0.031%

    No Known Activations