© 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. 14664
    Prev
    Next
    INDEX
    Explanations

    words related to criminal offenses, particularly those involving sexual misconduct

    oai_token-act-pair · gpt-3.5-turbo

    terms related to sexual offenses involving minors

    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
    DragonMagazine
    -0.79
    ĪĴ
    -0.79
     Defenders
    -0.75
    hani
    -0.75
    PRESS
    -0.66
    winner
    -0.65
    SHIP
    -0.63
    BOOK
    -0.62
    fight
    -0.62
     laureate
    -0.62
    POSITIVE LOGITS
    estation
    1.16
    asses
    1.02
    reated
    0.96
    onte
    0.92
    eties
    0.91
    icka
    0.91
    esting
    0.91
     mol
    0.90
    osexual
    0.88
    olit
    0.87
    Activations Density 0.009%

    No Known Activations