© 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. 4-RES-JB
    11. 12524
    Prev
    Next
    INDEX
    Explanations

    mentions of releases or reports

    oai_token-act-pair · gpt-3.5-turbo

    references to religious affiliations or contexts

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

    No Comments

    Negative Logits
    HAEL
    -0.92
    enegger
    -0.67
    BOX
    -0.67
    manship
    -0.66
    cules
    -0.65
    schild
    -0.65
    terday
    -0.64
    olicy
    -0.63
     Pigs
    -0.62
    ãģ®ç
    -0.62
    POSITIVE LOGITS
    iance
    1.19
    iability
    1.19
    iever
    1.17
    igion
    1.16
    iable
    1.06
    atively
    1.06
    inqu
    1.00
    ieved
    0.99
    ievers
    0.97
    atives
    0.96
    Activations Density 0.009%

    No Known Activations