© 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. 9-RES-JB
    11. 1476
    Prev
    Next
    INDEX
    Explanations

    phrases related to costs and expenses

    oai_token-act-pair · gpt-3.5-turbo

    phrases related to being "out of control" or associated with limitations and restrictions

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 9-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    9
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Cosponsors
    -0.77
    reddits
    -0.73
    _-
    -0.72
    illac
    -0.70
    ancial
    -0.67
    phal
    -0.65
    asks
    -0.65
    oÄŁ
    -0.64
     Ranking
    -0.63
    successful
    -0.61
    POSITIVE LOGITS
     bounds
    1.01
     nowhere
    0.83
     sync
    0.78
    sync
    0.75
     sight
    0.74
     reach
    0.71
     hiber
    0.70
     kil
    0.68
    town
    0.68
    wed
    0.67
    Activations Density 0.042%

    No Known Activations