© 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. 3-RES-JB
    11. 15719
    Prev
    Next
    INDEX
    Explanations

    phrases or terms related to illegal drugs, specifically crack cocaine

    oai_token-act-pair · gpt-3.5-turbo

    mentions of the word "crack" in various contexts

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 3-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    3
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ikk
    -0.89
     Leth
    -0.76
     Pwr
    -0.72
     Fenrir
    -0.67
    urse
    -0.61
    DK
    -0.60
     dracon
    -0.59
     traged
    -0.59
     Constantin
    -0.59
    velength
    -0.58
    POSITIVE LOGITS
    pots
    1.10
    ible
    1.02
    Berry
    0.93
    lings
    0.91
    ibly
    0.87
    pot
    0.87
    er
    0.86
    ers
    0.83
    vine
    0.80
     cocaine
    0.79
    Activations Density 0.007%

    No Known Activations