© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 11-RES-JB
    11. 21980
    Prev
    Next
    INDEX
    Explanations

    phrases related to escalation or potential negative consequences

    oai_token-act-pair · gpt-3.5-turbo

    phrases related to risk, danger, and consequences

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 11-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    11
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    atorium
    -0.65
     ado
    -0.62
    eatures
    -0.60
    rians
    -0.58
     salute
    -0.58
    raints
    -0.57
    \":
    -0.56
    ilib
    -0.56
    ories
    -0.55
     remembrance
    -0.55
    POSITIVE LOGITS
     depending
    1.38
     if
    1.27
    depending
    1.24
    if
    1.01
     someday
    0.97
     anytime
    0.84
     IF
    0.84
     easily
    0.82
     unless
    0.80
    sometimes
    0.78
    Activations Density 0.433%

    No Known Activations