context-compressor
Public
bench / results / 2026-08-24-deepseek-r1-distill-llama-8b.md
bench / results / 2026-08-24-deepseek-r1-distill-llama-8b.md
Model: deepseek-r1-distill-llama-8b
Settings: {"limitTokens":500,"floorTokens":400,"keepRecentTokens":300,"chunkTokens":800,"summarizerMaxTokens":400,"consolidationBudgetTokens":600,"consolidationKeepNewest":2,"mergeMaxTokens":400,"maxExcerptChars":32000,"quizMaxTokens":120}
| Fixture | Recall (compacted) | Recall (consolidated) | Reduction | Chunk latency mean/p50/max (ms) | Structure | Cache |
|---|---|---|---|---|---|---|
| coding-agent | 3/5 | — | 42% | 15771.3/15226/17148 | ✅ | ✅ |
| research | 0/3 | — | 46% | 12293/12293/12293 | ✅ | ✅ |
| contradictory-updates | 3/3 | — | 9% | 15196/15196/15196 | ✅ | ✅ |
| tool-spam | 3/3 | 3/3 | 43% | 15885.8/13929/53297 | ✅ | ✅ |
Recall = seeded facts answered correctly from the compacted view alone. Results are comparable only within one model; see bench/README.md.
Model: deepseek-r1-distill-llama-8b
Settings: {"limitTokens":500,"floorTokens":400,"keepRecentTokens":300,"chunkTokens":800,"summarizerMaxTokens":400,"consolidationBudgetTokens":600,"consolidationKeepNewest":2,"mergeMaxTokens":400,"maxExcerptChars":32000,"quizMaxTokens":120}
| Fixture | Recall (compacted) | Recall (consolidated) | Reduction | Chunk latency mean/p50/max (ms) | Structure | Cache |
|---|---|---|---|---|---|---|
| coding-agent | 3/5 | — | 42% | 15771.3/15226/17148 | ✅ | ✅ |
| research | 0/3 | — | 46% | 12293/12293/12293 | ✅ | ✅ |
| contradictory-updates | 3/3 | — | 9% | 15196/15196/15196 | ✅ | ✅ |
| tool-spam | 3/3 | 3/3 | 43% | 15885.8/13929/53297 | ✅ | ✅ |
Recall = seeded facts answered correctly from the compacted view alone. Results are comparable only within one model; see bench/README.md.