© 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. 11-RES-JB
    11. 20407
    Prev
    Next
    INDEX
    Explanations

    personal interactions and decisions involving people in different roles or positions

    oai_token-act-pair · gpt-3.5-turbo

    instances of personal pronouns and their relationships in the context of dialogue or interpersonal interactions

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 11-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    11
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Survivors
    -0.66
     diagnostic
    -0.65
     snapshots
    -0.64
     Illum
    -0.63
    ãĤ¹ãĥĪ
    -0.62
     Diagn
    -0.61
    Memory
    -0.60
     havoc
    -0.59
     diagn
    -0.59
    diagn
    -0.59
    POSITIVE LOGITS
     agreed
    1.24
     accepted
    1.19
     complied
    1.17
     reluctantly
    1.13
     acquies
    1.13
     gladly
    1.12
     refused
    1.10
     balk
    1.08
     declined
    1.07
     rejected
    1.02
    Activations Density 0.553%

    No Known Activations