© 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. 2847
    Prev
    Next
    INDEX
    Explanations

    strong negative sentiments or expressions of disapproval towards certain topics or individuals

    oai_token-act-pair · gpt-3.5-turbo

    sentiments of dislike or distrust towards people, topics, or situations

    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
    Not in Any Lists

    No Comments

    Negative Logits
     Olymp
    -0.74
    ortun
    -0.73
    itialized
    -0.72
     timing
    -0.71
    lymp
    -0.71
     timely
    -0.69
    Plex
    -0.67
     motor
    -0.67
    cod
    -0.67
     lucky
    -0.65
    POSITIVE LOGITS
     dislike
    2.53
     disliked
    2.42
     despise
    2.03
     disdain
    1.99
     distrust
    1.95
     disapprove
    1.94
     despised
    1.94
     scorn
    1.90
     hated
    1.88
     disapproval
    1.77
    Activations Density 0.049%

    No Known Activations