Pith. sign in

REVIEW 5 cited by

Learning to Compress Prompts with Gist Tokens

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2304.08467 v3 pith:XQWYELDC submitted 2023-04-17 cs.CL

classification cs.CL
keywords promptsgistcompresscompressionfinetuninggistingmodelsprompt
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

Prompting is the primary way to utilize the multitask capabilities of language models (LMs), but prompts occupy valuable space in the input context window, and repeatedly encoding the same prompt is computationally inefficient. Finetuning and distillation methods allow for specialization of LMs without prompting, but require retraining the model for each task. To avoid this trade-off entirely, we present gisting, which trains an LM to compress prompts into smaller sets of "gist" tokens which can be cached and reused for compute efficiency. Gist models can be trained with no additional cost over standard instruction finetuning by simply modifying Transformer attention masks to encourage prompt compression. On decoder (LLaMA-7B) and encoder-decoder (FLAN-T5-XXL) LMs, gisting enables up to 26x compression of prompts, resulting in up to 40% FLOPs reductions, 4.2% wall time speedups, and storage savings, all with minimal loss in output quality.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 5 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. OpenAlex reports about 17 citations worldwide. Full citation record

  1. What to Keep, What to Forget: A Rate--Distortion View of Memory Compaction in LLMs and Agents

    cs.LG 2026-07 conditional novelty 6.0 of 10

    KV-cache eviction, prompt compression, recurrent state bounding, and agent memory consolidation are unified as one rate-distortion problem with a shared lower bound, shared failure mode, and transferable mechanisms.

  2. ARC-Encoder: learning compressed text representations for large language models

    cs.CL 2025-10 conditional novelty 6.0 of 10

    ARC-Encoder pools queries in an encoder's last attention layer to produce compressed continuous representations that a frozen decoder consumes as token embeddings.

  3. GEM: Empowering LLM for both Embedding Generation and Language Understanding

    cs.CL 2025-06 conditional novelty 6.0 of 10

    GEM fine-tunes decoder-only LLMs with bottleneck special tokens and a mixed next-token prediction and contrastive objective, giving them text-embedding ability with only a modest MMLU drop.

  4. Token-Operations-Oriented Inference Optimization Techniques for Large Models

    cs.SE 2026-06 unverdicted novelty 3.0 of 10

    The paper introduces a four-layer technical architecture for token-operations-oriented inference optimization in large models and reviews key technologies and industry status at each layer.

  5. Breaking Thought Patterns: A Multi-Dimensional Reasoning Framework for LLMs

    cs.CL 2025-06 reject novelty 2.0 of 10

    LADDER, a proposed mix of chain-of-thought prompting, mixture-of-experts layers, and linear projections, reportedly improves LLM creativity and diversity, but the evidence is thin and partly contradictory.

Pith tools