LatComm compresses LLM context once for reuse across models
On QuALITY the receiver sees 57 times less input and answers an article's questions three times faster.
Stories per day · 30 days
Relatedbenchmarks2agent-memory1agents1coding-agents1cross-attention1direct-preference-optimization1faithfulness1file-selection1keep-bench1latent-communication1llm-agents1long-context1long-horizon1question-answering1qwen1retrieval1swe-bench1tool-use1
5 stories, newest first · one card per story
Stories tagged #context-compression
KEEP-Bench finds context compressors often drop required evidence
The best open-weight 31B model keeps all required evidence in 66.8% of 4,102 cases, and the best proprietary model in 61.6%.
Coding agents fix more issues when shown whole files
On 300 SWE-rebench issues, verbatim file selection beats compressors by 6.3 points with either model.
DeLoop cuts repeated tool-call loops in long-horizon agents
Under CodeGym compression, Qwen3-8B tasks containing loops fall from 57.3% to 6.1%.
FlowState keeps execution state as memory for long-horizon agents
Against a full-context DeepSeek-V4-Flash baseline, authors report higher pass rates and about 40% fewer tokens.