© 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. 0-RES-JB
    11. 546
    Prev
    Next
    INDEX
    Explanations

    words related to sadness or unfortunate situations

    oai_token-act-pair · gpt-3.5-turbo

    expressions of sadness or regret

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 0-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    0
    Activation Function
    relu
    Embeds
    IFrame
    Link
    •layer 0 - sadness

    No Comments

    Negative Logits
     initiation
    -0.79
    ized
    -0.71
    atum
    -0.67
    NRS
    -0.66
     interference
    -0.65
     Impact
    -0.64
    imeter
    -0.64
     initiating
    -0.64
    -,
    -0.64
     depos
    -0.64
    POSITIVE LOGITS
     sadly
    3.05
    Sadly
    2.04
     Sadly
    1.94
     tragically
    1.89
     sorely
    1.80
     alas
    1.67
     thankfully
    1.54
     curiously
    1.36
     Alas
    1.32
     strangely
    1.29
    Activations Density 0.028%

    No Known Activations