© 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. 2-RES-JB
    11. 13464
    Prev
    Next
    INDEX
    Explanations

    words related to extreme negative situations or conditions

    oai_token-act-pair · gpt-3.5-turbo

    expressions indicating the concept of "worst" in various contexts

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 2-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    2
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    arij
    -0.79
    bara
    -0.76
    alde
    -0.72
    ependence
    -0.72
    paio
    -0.72
    ighters
    -0.71
    urches
    -0.70
    iverpool
    -0.70
    arya
    -0.70
    dinand
    -0.70
    POSITIVE LOGITS
     nightmare
    1.01
     worst
    0.95
    worst
    0.93
     Worst
    0.91
     imaginable
    0.89
     offender
    0.89
     offenders
    0.85
     loser
    0.84
     losers
    0.84
     nightmares
    0.81
    Activations Density 0.008%

    No Known Activations