© 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. 8-RES-JB
    11. 16733
    Prev
    Next
    INDEX
    Explanations

    negative statements or sentiments

    oai_token-act-pair · gpt-3.5-turbo

    negative expressions or refusal statements

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    Offline
    -0.72
     Nep
    -0.63
    Edge
    -0.60
     Northwest
    -0.60
    advertisement
    -0.60
    Vers
    -0.60
     skept
    -0.59
    ynchron
    -0.57
    Prot
    -0.56
     Timbers
    -0.55
    POSITIVE LOGITS
    plete
    0.86
     nor
    0.84
    existent
    0.82
    isi
    0.74
    etheless
    0.74
    nor
    0.73
    atel
    0.73
    obin
    0.73
    itto
    0.70
     malice
    0.68
    Activations Density 0.279%

    No Known Activations