cd/entity/Prompt Compression and Contrastive Conditioning for Controllability and Toxicity Reduction in Language Models· home› entities› Prompt Compression and Contrastive Conditioning for Controllability and Toxicity Reduction in Language Models

grep -l @prompt compression and contrastive conditioning for controllability and toxicity reduction in language models /news/*.json | wc -l → 1

Prompt Compression and Contrastive Conditioning for Controllability and Toxicity Reduction in Language Models

mentions 1 type Person feed RSS

// recent coverage 1 mentions

01:42

2026-08-21

shopify.engineering

large-language-models

Gisting: Compressing LLM Agent context to ↑ throughput and ↓ cost

Shopify Engineering implemented gisting to compress the Sidekick GraphQL agent's system prompt from ~6,000 tokens to ~1,500 gist tokens (a 4:1 reduction) without losing prediction quality, cutting med…

// co-occurs with top 2 entities

Shopify Engineering 1 Sidekick GraphQL agent 1

// topics top 4 topics

large language models 1 ai infrastructure 1 ai research 1 natural language processing 1