Compress the Context, Keep the Commitments: A Formal Framework for Verifiable LLM Context Compression
For LLM developers and researchers, this provides a principled, verifiable approach to context compression, addressing the lack of commitment-preserving methods in existing systems.
The paper introduces Context Codec, a formal framework for LLM context compression that preserves semantic commitments (goals, constraints, decisions) rather than just tokens. It defines metrics like Critical Atom Recall and a compact rendering language (CCL), showing in a small study that CCL-Core balances explicitness and compactness between prose and JSON.
LLM context is not just tokens; it is a set of commitments. Long-running conversations accumulate goals, constraints, decisions, preferences, tool results, retrieved evidence, artifacts, and safety boundaries that future responses must preserve. Existing context-management methods reduce length through truncation, retrieval, summarization, memory systems, or token-level prompt compression, but they rarely specify which semantic commitments must survive compression or how their preservation should be measured. We propose Context Codec, a commitment-level framework for compressing prompts and chat histories. Context Codec represents dialogue state as typed, source-grounded semantic atoms with canonical identity, equivalence, conflict, confidence, risk, and evidence spans. It separates five concerns - extraction, normalization, representation, rendering, and verification - and introduces metrics for Critical Atom Recall, Weighted Atom Recall, Commitment Density, and round-trip recoverability. It also defines a taxonomy of semantic compression errors, a concrete normalization procedure, conservative fallback rules for low-confidence and safety-critical atoms, and Context Compression Language (CCL), an ASCII-first compact rendering of canonical JSON atoms. In a small diagnostic study, CCL-Core occupies a useful middle ground between structured prose and JSON: more explicit and auditable than prose, usually more compact than JSON, and less risky than heavily minified notation. The result is not a claim that shorthand solves compression, but a framework for making context compression verifiable: compress the conversation, keep the commitments.