© 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. 863
    Prev
    Next
    INDEX
    Explanations

    legal or criminal terms related to betrayal or misconduct

    oai_token-act-pair · gpt-3.5-turbo

    keywords related to legal crimes and offenses

    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
    izen
    -0.76
    mega
    -0.76
    aird
    -0.71
    illac
    -0.71
    iman
    -0.70
    liner
    -0.69
    ixed
    -0.68
    ãĤ©
    -0.68
     Ott
    -0.67
    este
    -0.67
    POSITIVE LOGITS
    ionage
    1.02
     accusations
    0.99
     accusation
    0.99
     charges
    0.98
     allegations
    0.96
     prevention
    0.92
     manslaughter
    0.89
     prosecutions
    0.87
     offences
    0.87
     allegation
    0.86
    Activations Density 0.140%

    No Known Activations