© 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. 543
    Prev
    Next
    INDEX
    Explanations

    references to basketball

    oai_token-act-pair · gpt-3.5-turbo

    references to basketball and wrestling

    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
     rud
    -0.69
     Rey
    -0.69
     rescue
    -0.68
     tac
    -0.65
    secondary
    -0.64
     intrusion
    -0.64
     descript
    -0.63
    erb
    -0.63
     nod
    -0.63
     ranger
    -0.62
    POSITIVE LOGITS
     Basketball
    3.34
    NBA
    1.39
     Tennis
    1.22
     Wrestling
    1.22
     Bullets
    1.20
    asketball
    1.15
     Mahjong
    1.13
     Boxing
    1.11
    BALL
    1.04
     Baseball
    0.96
    Activations Density 0.032%

    No Known Activations