© 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 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. 20041
    Prev
    Next
    INDEX
    Explanations

    words related to stopping or halting action

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "halt" and its variations

    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
     Dynamics
    -0.80
    Sov
    -0.76
    rics
    -0.76
    ãĤĮ
    -0.72
    çīĪ
    -0.71
    iosyncr
    -0.67
    ographies
    -0.65
    ãģĦ
    -0.64
    ramid
    -0.63
    ortium
    -0.62
    POSITIVE LOGITS
     halt
    1.01
    steen
    0.89
     halting
    0.80
    downs
    0.79
     derail
    0.79
    ŃĶ
    0.78
     seiz
    0.77
     halted
    0.76
    ength
    0.76
    down
    0.75
    Activations Density 0.016%

    No Known Activations