# Inter-Context Knowledge Conflict on PopQA

Source: [Inter-Context Knowledge Conflict on PopQA](<https://www.sota2.com/research/sota/inter-context-knowledge-conflict-on-popqa>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-07-01T07:46:48.218Z

Evaluation of model performance on the PopQA dataset focusing on Inter-Context Knowledge Conflict resolution, measured using ACCL, ACCEM, and ACCCEM.

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

| Scope | Recorded value |
| --- | --- |
| Task | Inter-Context Knowledge Conflict |
| Dataset (including any stated split) | popQA |
| Metric | ACCL |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 32 of 32 |
| Coverage | All results currently returned for this benchmark |

Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol.

## Metrics

- [ACCL](<https://www.sota2.com/research/sota/inter-context-knowledge-conflict-on-popqa.md?metric=ACCL>): Higher is better; unit: PERCENT.

- [ACCEM](<https://www.sota2.com/research/sota/inter-context-knowledge-conflict-on-popqa.md?metric=ACCEM>): Higher is better; unit: PERCENT.

- [ACCCEM](<https://www.sota2.com/research/sota/inter-context-knowledge-conflict-on-popqa.md?metric=ACCCEM>): Higher is better; unit: PERCENT.

## Recorded results

| Method | ACCL | Result date | Report paper |
| --- | --- | --- | --- |
| KCR (“Graph”) | 90.28 | Not recorded | 68fa7a621e209131cea82c1f |
| KCR (“Graph”) | 87.5 | Not recorded | 68fa7a621e209131cea82c1f |
| KCR (“Text”) | 86.67 | Not recorded | 68fa7a621e209131cea82c1f |
| KCR (“Text”) | 84.17 | Not recorded | 68fa7a621e209131cea82c1f |
| KCR (“Graph”) | 80.14 | Not recorded | 68fa7a621e209131cea82c1f |
| KCR (“Text”) | 74.72 | Not recorded | 68fa7a621e209131cea82c1f |
| KCR (“Graph”) | 69.58 | Not recorded | 68fa7a621e209131cea82c1f |
| KCR (“Text”) | 68.61 | Not recorded | 68fa7a621e209131cea82c1f |
| ASTUTE RAG (GPT-5.1) | 66.94 | Not recorded | 68fa7a621e209131cea82c1f |
| Llama 3.1-8B | 56.53 | Not recorded | 68fa7a621e209131cea82c1f |
| ASTUTE RAG (GPT-4o) | 54.86 | Not recorded | 68fa7a621e209131cea82c1f |
| Directly Use Backbone | 54.31 | Not recorded | 68fa7a621e209131cea82c1f |
| Directly Use Backbone | 54.31 | Not recorded | 68fa7a621e209131cea82c1f |
| Qwen 2.5-3B | 51.67 | Not recorded | 68fa7a621e209131cea82c1f |
| ASTUTE RAG | 50.28 | Not recorded | 68fa7a621e209131cea82c1f |
| Directly Use Backbone | 49.58 | Not recorded | 68fa7a621e209131cea82c1f |
| GPT-5.1 (Thinking) | 49.44 | Not recorded | 68fa7a621e209131cea82c1f |
| Directly Use Backbone | 48.47 | Not recorded | 68fa7a621e209131cea82c1f |
| ASTUTE RAG | 43.75 | Not recorded | 68fa7a621e209131cea82c1f |
| CAD Decoding | 40.28 | Not recorded | 68fa7a621e209131cea82c1f |
| CAD Decoding | 40.28 | Not recorded | 68fa7a621e209131cea82c1f |
| Qwen 2.5-1.5B | 39.86 | Not recorded | 68fa7a621e209131cea82c1f |
| ASTUTE RAG | 39.72 | Not recorded | 68fa7a621e209131cea82c1f |
| CUAD (Default) | 34.31 | Not recorded | 68fa7a621e209131cea82c1f |
| ADACAD (Default) | 34.31 | Not recorded | 68fa7a621e209131cea82c1f |
| COCOA (Default) | 34.31 | Not recorded | 68fa7a621e209131cea82c1f |
| GPT-4o-mini | 30.42 | Not recorded | 68fa7a621e209131cea82c1f |
| Llama 3.2-1B | 30 | Not recorded | 68fa7a621e209131cea82c1f |
| CAD Decoding | 29.72 | Not recorded | 68fa7a621e209131cea82c1f |
| Llama 3.2-3B | 17.64 | Not recorded | 68fa7a621e209131cea82c1f |
| Qwen 3-8B (Thinking) | 17.36 | Not recorded | 68fa7a621e209131cea82c1f |
| Qwen 2.5-7B | 12.5 | Not recorded | 68fa7a621e209131cea82c1f |

### KCR (“Graph”)

ACCL: 90.28. Result ID: 6a44c5e773d097d845beb6dc.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a621e209131cea82c1f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7a621e209131cea82c1f.md>)

Recorded evaluation configuration: {"Backbone":"Qwen 3-8B","Format":"Graph"}. This record does not assert independent reproduction.

### KCR (“Graph”)

ACCL: 87.5. Result ID: 6a44c5e773d097d845beb6cf.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a621e209131cea82c1f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7a621e209131cea82c1f.md>)

Recorded evaluation configuration: {"Backbone":"Llama 3.2-3B","Format":"Graph"}. This record does not assert independent reproduction.

### KCR (“Text”)

ACCL: 86.67. Result ID: 6a44c5e773d097d845beb6d3.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a621e209131cea82c1f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7a621e209131cea82c1f.md>)

Recorded evaluation configuration: {"Backbone":"Qwen 2.5-7B","Format":"Text"}. This record does not assert independent reproduction.

### KCR (“Text”)

ACCL: 84.17. Result ID: 6a44c5e773d097d845beb6ce.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a621e209131cea82c1f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7a621e209131cea82c1f.md>)

Recorded evaluation configuration: {"Backbone":"Llama 3.2-3B","Format":"Text"}. This record does not assert independent reproduction.

### KCR (“Graph”)

ACCL: 80.14. Result ID: 6a44c5e773d097d845beb6d4.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a621e209131cea82c1f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7a621e209131cea82c1f.md>)

Recorded evaluation configuration: {"Backbone":"Qwen 2.5-7B","Format":"Graph"}. This record does not assert independent reproduction.

### KCR (“Text”)

ACCL: 74.72. Result ID: 6a44c5e773d097d845beb6db.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a621e209131cea82c1f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7a621e209131cea82c1f.md>)

Recorded evaluation configuration: {"Backbone":"Qwen 3-8B","Format":"Text"}. This record does not assert independent reproduction.

### KCR (“Graph”)

ACCL: 69.58. Result ID: 6a44c5e773d097d845beb6d7.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a621e209131cea82c1f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7a621e209131cea82c1f.md>)

Recorded evaluation configuration: {"Backbone":"Qwen 3-8B (Thinking)","Format":"Graph"}. This record does not assert independent reproduction.

### KCR (“Text”)

ACCL: 68.61. Result ID: 6a44c5e773d097d845beb6d6.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a621e209131cea82c1f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7a621e209131cea82c1f.md>)

Recorded evaluation configuration: {"Backbone":"Qwen 3-8B (Thinking)","Format":"Text"}. This record does not assert independent reproduction.

### ASTUTE RAG (GPT-5.1)

ACCL: 66.94. Result ID: 6a44c5e773d097d845beb6ca.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a621e209131cea82c1f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79061e209131cea7bc2c.md>)

Recorded evaluation configuration: {"Model Family":"General Baselines (With Context)"}. This record does not assert independent reproduction.

### Llama 3.1-8B

ACCL: 56.53. Result ID: 6a44c5e773d097d845beb6c3.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a621e209131cea82c1f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6eae1e209131cea41685.md>)

[Implementation](<https://github.com/facebookresearch/llama>)

Recorded evaluation configuration: {"Model Family":"General Baselines (With Context)"}. This record does not assert independent reproduction.

### ASTUTE RAG (GPT-4o)

ACCL: 54.86. Result ID: 6a44c5e773d097d845beb6c9.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a621e209131cea82c1f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79061e209131cea7bc2c.md>)

Recorded evaluation configuration: {"Model Family":"General Baselines (With Context)"}. This record does not assert independent reproduction.

### Directly Use Backbone

ACCL: 54.31. Result ID: 6a44c5e773d097d845beb6d5.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a621e209131cea82c1f.md>)

Recorded evaluation configuration: {"Backbone":"Qwen 3-8B (Thinking)"}. This record does not assert independent reproduction.

### Directly Use Backbone

ACCL: 54.31. Result ID: 6a44c5e773d097d845beb6d8.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a621e209131cea82c1f.md>)

Recorded evaluation configuration: {"Backbone":"Qwen 3-8B"}. This record does not assert independent reproduction.

### Qwen 2.5-3B

ACCL: 51.67. Result ID: 6a44c5e773d097d845beb6c5.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a621e209131cea82c1f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa767e1e209131cea6d6a7.md>)

Recorded evaluation configuration: {"Model Family":"General Baselines (With Context)"}. This record does not assert independent reproduction.

### ASTUTE RAG

ACCL: 50.28. Result ID: 6a44c5e773d097d845beb6da.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a621e209131cea82c1f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79061e209131cea7bc2c.md>)

Recorded evaluation configuration: {"Backbone":"Qwen 3-8B"}. This record does not assert independent reproduction.

### Directly Use Backbone

ACCL: 49.58. Result ID: 6a44c5e773d097d845beb6cb.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a621e209131cea82c1f.md>)

Recorded evaluation configuration: {"Backbone":"Llama 3.2-3B"}. This record does not assert independent reproduction.

### GPT-5.1 (Thinking)

ACCL: 49.44. Result ID: 6a44c5e773d097d845beb6c1.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a621e209131cea82c1f.md>)

Recorded evaluation configuration: {"Model Family":"Direct Generation (Without Context)"}. This record does not assert independent reproduction.

### Directly Use Backbone

ACCL: 48.47. Result ID: 6a44c5e773d097d845beb6d0.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a621e209131cea82c1f.md>)

Recorded evaluation configuration: {"Backbone":"Qwen 2.5-7B"}. This record does not assert independent reproduction.

### ASTUTE RAG

ACCL: 43.75. Result ID: 6a44c5e773d097d845beb6d2.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a621e209131cea82c1f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79061e209131cea7bc2c.md>)

Recorded evaluation configuration: {"Backbone":"Qwen 2.5-7B"}. This record does not assert independent reproduction.

### CAD Decoding

ACCL: 40.28. Result ID: 6a44c5e773d097d845beb6d1.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a621e209131cea82c1f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6f6e1e209131cea45813.md>)

Recorded evaluation configuration: {"Backbone":"Qwen 2.5-7B"}. This record does not assert independent reproduction.

### CAD Decoding

ACCL: 40.28. Result ID: 6a44c5e773d097d845beb6d9.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a621e209131cea82c1f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6f6e1e209131cea45813.md>)

Recorded evaluation configuration: {"Backbone":"Qwen 3-8B"}. This record does not assert independent reproduction.

### Qwen 2.5-1.5B

ACCL: 39.86. Result ID: 6a44c5e773d097d845beb6c4.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a621e209131cea82c1f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa767e1e209131cea6d6a7.md>)

Recorded evaluation configuration: {"Model Family":"General Baselines (With Context)"}. This record does not assert independent reproduction.

### ASTUTE RAG

ACCL: 39.72. Result ID: 6a44c5e773d097d845beb6cd.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a621e209131cea82c1f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79061e209131cea7bc2c.md>)

Recorded evaluation configuration: {"Backbone":"Llama 3.2-3B"}. This record does not assert independent reproduction.

### CUAD (Default)

ACCL: 34.31. Result ID: 6a44c5e773d097d845beb6c6.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a621e209131cea82c1f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73131e209131cea5a54b.md>)

Recorded evaluation configuration: {"Model Family":"General Baselines (With Context)"}. This record does not assert independent reproduction.

### ADACAD (Default)

ACCL: 34.31. Result ID: 6a44c5e773d097d845beb6c7.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a621e209131cea82c1f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78411e209131cea776f7.md>)

Recorded evaluation configuration: {"Model Family":"General Baselines (With Context)"}. This record does not assert independent reproduction.

### COCOA (Default)

ACCL: 34.31. Result ID: 6a44c5e773d097d845beb6c8.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a621e209131cea82c1f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7add1e209131cea8594e.md>)

Recorded evaluation configuration: {"Model Family":"General Baselines (With Context)"}. This record does not assert independent reproduction.

### GPT-4o-mini

ACCL: 30.42. Result ID: 6a44c5e773d097d845beb6c0.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a621e209131cea82c1f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa727a1e209131cea56a1e.md>)

[Implementation](<https://github.com/openai/evals>)

Recorded evaluation configuration: {"Model Family":"Direct Generation (Without Context)"}. This record does not assert independent reproduction.

### Llama 3.2-1B

ACCL: 30. Result ID: 6a44c5e773d097d845beb6c2.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a621e209131cea82c1f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6eae1e209131cea41685.md>)

[Implementation](<https://github.com/facebookresearch/llama>)

Recorded evaluation configuration: {"Model Family":"General Baselines (With Context)"}. This record does not assert independent reproduction.

### CAD Decoding

ACCL: 29.72. Result ID: 6a44c5e773d097d845beb6cc.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a621e209131cea82c1f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6f6e1e209131cea45813.md>)

Recorded evaluation configuration: {"Backbone":"Llama 3.2-3B"}. This record does not assert independent reproduction.

### Llama 3.2-3B

ACCL: 17.64. Result ID: 6a44c5e773d097d845beb6bd.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a621e209131cea82c1f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6eae1e209131cea41685.md>)

[Implementation](<https://github.com/facebookresearch/llama>)

Recorded evaluation configuration: {"Model Family":"Direct Generation (Without Context)"}. This record does not assert independent reproduction.

### Qwen 3-8B (Thinking)

ACCL: 17.36. Result ID: 6a44c5e773d097d845beb6bf.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a621e209131cea82c1f.md>)

Recorded evaluation configuration: {"Model Family":"Direct Generation (Without Context)"}. This record does not assert independent reproduction.

### Qwen 2.5-7B

ACCL: 12.5. Result ID: 6a44c5e773d097d845beb6be.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a621e209131cea82c1f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa767e1e209131cea6d6a7.md>)

Recorded evaluation configuration: {"Model Family":"Direct Generation (Without Context)"}. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.
