Datasets:
Update README.md
Browse files
README.md
CHANGED
|
@@ -4,9 +4,10 @@ language:
|
|
| 4 |
- en
|
| 5 |
---
|
| 6 |
# PI-LLM: The Core Retrieval Challenge Behind MRCR
|
| 7 |
-
Multi-round co-reference interference in long context
|
| 8 |
(ICML2025 Long-Context Workshop Accepted)
|
| 9 |
|
|
|
|
|
|
|
| 10 |
Classic long-context benchmarks often test retrieving a single "needle" from a massive "haystack." MRCR raises the bar by placing many similar needles in the same context, requiring models to select the correct item (up to 8 needles), and shows that all LLMs struggle with this task.
|
| 11 |
|
| 12 |
- OpenAI MRCR dataset: https://huggingface.co/datasets/openai/mrcr
|
|
|
|
| 4 |
- en
|
| 5 |
---
|
| 6 |
# PI-LLM: The Core Retrieval Challenge Behind MRCR
|
|
|
|
| 7 |
(ICML2025 Long-Context Workshop Accepted)
|
| 8 |
|
| 9 |
+
Multi-round co-reference in Context Interference:
|
| 10 |
+
|
| 11 |
Classic long-context benchmarks often test retrieving a single "needle" from a massive "haystack." MRCR raises the bar by placing many similar needles in the same context, requiring models to select the correct item (up to 8 needles), and shows that all LLMs struggle with this task.
|
| 12 |
|
| 13 |
- OpenAI MRCR dataset: https://huggingface.co/datasets/openai/mrcr
|