© 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. 10913
    Prev
    Next
    INDEX
    Explanations

    phrases related to controversial or incriminating information

    oai_token-act-pair · gpt-3.5-turbo

    references to documents and reports, particularly those related to accusations or controversies

    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
    rl
    -0.84
     Pyr
    -0.78
     Saw
    -0.75
    rh
    -0.71
    jc
    -0.71
    ahime
    -0.71
     Prairie
    -0.69
     Instrument
    -0.69
    itans
    -0.68
    ulton
    -0.66
    POSITIVE LOGITS
     dossier
    1.59
    ossier
    0.97
     Steele
    0.88
     leveled
    0.87
     Ukrain
    0.85
     blackmail
    0.82
     allegations
    0.81
     profiles
    0.76
     compiled
    0.76
     trove
    0.75
    Activations Density 0.020%

    No Known Activations