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

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI 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. 12-RES-JB
    11. 3131
    Prev
    Next
    INDEX
    Explanations

    phrases related to political discussions and international relations, particularly focusing on potential consequences and actions

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot

    conditional phrases that suggest potential outcomes or risks

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 12-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_post
    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_post
    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
    PDATE
    -1.14
     Seym
    -1.02
    anamo
    -0.99
    lication
    -0.88
     Cosponsors
    -0.84
     practition
    -0.81
    isSpecial
    -0.80
    ß
    -0.80
    ThumbnailImage
    -0.79
    ortunately
    -0.76
    POSITIVE LOGITS
     contam
    1.33
     endanger
    1.31
     escalate
    1.28
     derail
    1.24
     worsen
    1.20
     result
    1.20
     cripp
    1.20
     devast
    1.19
     provoke
    1.18
     undermine
    1.17
    Activations Density 0.162%

    No Known Activations