Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    3. GPT2-Small
    4. Residual Stream
    5. 3-RES-JB
    6. 1979
    Prev
    Next
    INDEX
    Explanations

    references to football player positions such as "WR" (Wide Receiver), "DB" (Defensive Back), and "OL" (Offensive Line)

    oai_token-act-pair · gpt-3.5-turbo

    positions and roles related to football players

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 3-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    3
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    20439
    -0.82
    netflix
    -0.74
     Newsp
    -0.71
    atom
    -0.70
    capital
    -0.69
    microsoft
    -0.68
     Debor
    -0.68
    sth
    -0.67
    consumer
    -0.65
     Orwell
    -0.64
    POSITIVE LOGITS
     Scouting
    0.96
     coach
    0.95
     linebacker
    0.95
     fullback
    0.93
    Coach
    0.92
     rookie
    0.91
     cornerback
    0.89
     guru
    0.88
     linebackers
    0.83
     Rookie
    0.81
    Activations Density 0.046%

    No Known Activations