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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & 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. 10-RES-JB
    6. 13131
    Prev
    Next
    INDEX
    Explanations

    phrases related to past states of being or situations that have changed over time

    oai_token-act-pair · gpt-3.5-turbo

    phrases indicating a transformation or transition from one state to another

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 10-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    10
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Retrieved
    -0.82
    2018
    -0.74
    wake
    -0.70
     Cosponsors
    -0.69
    update
    -0.69
     2018
    -0.68
     Extend
    -0.68
     Recall
    -0.67
    "]=>
    -0.66
     Whereas
    -0.66
    POSITIVE LOGITS
     unthinkable
    1.24
     taboo
    1.04
     unimaginable
    1.02
     innocuous
    1.00
     harmless
    0.97
     dormant
    0.96
     regarded
    0.92
    thinkable
    0.92
     unheard
    0.90
     timid
    0.87
    Activations Density 0.267%

    No Known Activations