© 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. 8-RES-JB
    11. 23402
    Prev
    Next
    INDEX
    Explanations

    references to legal and regulatory entities, such as Citizenship and Immigration Services, concealed pistol licenses, and tax code designations

    oai_token-act-pair · gpt-3.5-turbo

    references to legal and organizational designations related to immigration and citizenship

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Atmosp
    -0.71
    lest
    -0.71
     scenes
    -0.71
    riz
    -0.69
     Subtle
    -0.69
     louder
    -0.69
    ohm
    -0.69
    mire
    -0.68
    tone
    -0.67
    stars
    -0.67
    POSITIVE LOGITS
     eligibility
    1.49
     exemption
    1.45
     registration
    1.42
     exemptions
    1.41
     paperwork
    1.33
     visas
    1.31
     visa
    1.30
     statutory
    1.28
     exempt
    1.28
     waivers
    1.27
    Activations Density 1.783%

    No Known Activations