© 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. 22322
    Prev
    Next
    INDEX
    Explanations

    the word "Ambient."

    oai_token-act-pair · gpt-3.5-turbo

    references to ambient music

    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
     toget
    -0.72
     cookie
    -0.71
    xxxxxxxx
    -0.71
     slur
    -0.70
     expiration
    -0.69
     narrowing
    -0.69
    uler
    -0.69
     shred
    -0.66
     claws
    -0.65
     paw
    -0.64
    POSITIVE LOGITS
     Amb
    3.88
    Amb
    3.32
     amb
    1.64
    amb
    1.58
     Ambrose
    1.52
     ambient
    1.32
     Ambassador
    1.27
     ambush
    1.12
     ambul
    1.12
     Camb
    1.09
    Activations Density 0.026%

    No Known Activations