© 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. 22584
    Prev
    Next
    INDEX
    Explanations

    titles or phrases that include the word "Goes"

    oai_token-act-pair · gpt-3.5-turbo

    phrases that include the words "goes" and "going"

    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
    âĶĢ
    -0.73
     snap
    -0.69
    lings
    -0.62
     punct
    -0.60
    fts
    -0.59
     âĶ
    -0.58
     butcher
    -0.58
     decoration
    -0.56
    ls
    -0.55
    ield
    -0.55
    POSITIVE LOGITS
     Goes
    2.48
     Going
    2.33
    Going
    1.77
     Gone
    1.07
    went
    1.05
     Went
    0.99
     Gets
    0.96
     Came
    0.94
     Demand
    0.94
    wcsstore
    0.94
    Activations Density 0.034%

    No Known Activations