© 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. 20078
    Prev
    Next
    INDEX
    Explanations

    proper names with varying last names

    oai_token-act-pair · gpt-3.5-turbo

    proper names of individuals mentioned in the text

    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
     aur
    -0.93
    Alice
    -0.84
    Scot
    -0.81
     Alice
    -0.81
     bir
    -0.79
     surround
    -0.72
     mah
    -0.71
     serpent
    -0.70
    arat
    -0.70
     hen
    -0.68
    POSITIVE LOGITS
     Jacobs
    2.70
     Scha
    2.25
     Liu
    1.92
     Rollins
    1.52
     McH
    1.49
     Kimmel
    1.38
     Braun
    1.34
     Fallon
    1.34
     Wolfe
    1.32
     Crowley
    1.28
    Activations Density 0.046%

    No Known Activations