paper_id
string
claim_id
string
claim
string
label
string
caption
string
evi_type
string
evi_path
string
context
string
domain
string
use_context
string
operation
string
paper_path
string
detail_others
string
license_name
string
license_url
string
claim_id_pair
string
evi_path_original
string
2025.clpsych-1.1
val_tab_1507
The application of the majority voting algorithm improved the RMSE on average by about 30% from 1.61 to 1.12 (columns GPT-4 {}_{\text{avg}} and GPT-4 {}_{\text{maj}} in Table 2 ).
Refuted
Table 2: RMSE results for research questions Q2, Q3, and Q4. GPT-4 stands for the GPT-4 annotator, Human for the human reader-annotator, avg : average of five GPT-4 completions/human guesses, maj : majority vote of five GPT-4 completions/human guesses, conf : majority vote of five GPT-4 completions/human guesses with c...
table
tables_png/dev/val_tab_1507.png
nlp
no
Change the cell values
papers/dev/nlp_2025.clpsych-1.1.json
CC BY 4.0
http://creativecommons.org/licenses/by/4.0/
0210
tables/dev/val_tab_1507.tex
2024.eacl-long.20
val_tab_1508
Second, pairwise comparisons with the baseline show both CG and finetuning are effective in formality control, where CG has slightly higher win ratio than FT against the baseline.
Refuted
Table 6: Human evaluation on Bengali, with quality on a 5-point scale \uparrow and formality on a 3-point scale ( \uparrow : formal) with standard deviations . Last two columns show pairwise comparison of formality scores to baseline NLLB-200 given the same source sentences (winning: scoring more in the direction of th...
table
tables_png/dev/val_tab_1508.png
The results are in Table 6 . First, adding attribute control does not appear to impact translation quality.
nlp
no
Change the cell values
papers/dev/nlp_2024.eacl-long.20.json
CC BY 4.0
https://creativecommons.org/licenses/by/4.0/
0116
tables/dev/val_tab_1508.tex
2024.eacl-long.12
val_tab_1509
Development sets are 5k instances samples from each set.
Refuted
Table 1: Dataset statistics for synthetic data generated in this work. We omit the average length of answers for fact verification as it is a classification task. SQ=Single-Query. TQ=Two-Queries.
table
tables_png/dev/val_tab_1509.png
We synthesize approximately 1.5 million multi-hop questions and 1.9 million claims. We use neucleus sampling (Holtzman et al., 2020 ) with a top- p probability of 0.9 for decoding when generating the data.
nlp
no
Change the cell values
papers/dev/nlp_2024.eacl-long.12.json
CC BY 4.0
https://creativecommons.org/licenses/by/4.0/
0155
tables/dev/val_tab_1509.tex
2410.17599
val_fig_1506
The underfitting delta model had a win rate of 54 when \alpha was 1.0, which increased to 64 when \alpha was 2.0.
Refuted
(a) Impact on instruction tuning; (b) Impact on unlearning; Impact of strength coefficient \alpha on performance
figure
figures/dev/val_fig_1506.png
For instruction tuning, we tested values within the range of [0.5, 2] and evaluated them on the first 50 data points of AlpacaEval. As shown in Figure 4(a) , we found that the performance was optimal when the \alpha value was 1.0, and increasing or decreasing \alpha resulted in decreased performance. This is similar to...
ml
yes
Legend Swap
papers/dev/ml_2410.17599.json
CC BY 4.0
http://creativecommons.org/licenses/by/4.0/
0039
null
2024.eacl-long.14
val_fig_1508
We observe that all languages that are included in the pre-training corpus of XLM-R, the cross-lingual transfer performs similarly to fully supervised methods.
Refuted
Figure 5: Comparison of Various Scenarios. We group languages by whether they and their scripts are seen in the pre-training corpus of XLM-R. Languages are ordered by the XLM-R fully-supervised performance in every group.
figure
figures/dev/val_fig_1508.png
Here, we compare the performance between cross-lingual transfer and fully-supervised methods.
nlp
no
Legend Swap
papers/dev/nlp_2024.eacl-long.14.json
CC BY 4.0
http://creativecommons.org/licenses/by/4.0/
0253
null
2023.ijcnlp-main.36
val_tab_1510
While the reported accuracy on the original test dataset is 38% for QAGNN and 40% for BioLinkBert, the accuracy on our 100 randomly selected demographic-independent questions use to construct the vignettes is 40% for QAGNN and 39% for BioLinkBert.
Refuted
Table 4: Accuracy (in percentages) of the two models on our demographically enhanced datasets. M =male; F =female; W =White; B =Black; A-A =African-American; H =Hispanic; As =Asian; SOr =sexual orientation; O*=original test dataset; O=the original, unmodified 100 vignettes; D=No demographic information; Gen=Gender; 1=Q...
table
tables_png/dev/val_tab_1510.png
nlp
no
Change the cell values
papers/dev/nlp_2023.ijcnlp-main.36.json
CC BY 4.0
http://creativecommons.org/licenses/by/4.0/
0123
tables/dev/val_tab_1510.tex
2025.naacl-long.13
val_tab_1511
For Factuality, the annotators demonstrate moderate agreement (Krippendorff’s Alpha = 0.68), suggesting that the conversations are consistently viewed as highly factual, which implies that the SQL queries in our dataset are of high quality.
Refuted
Table 4: Summary of Human Annotation Scores for Naturalness, Factuality, and Helpfulness.
table
tables_png/dev/val_tab_1511.png
Table 4 shows the mean, standard deviation, and Krippendorff’s Alpha for inter-annotator agreement. The high mean scores close to 1 (1.15-1.5) and substantial agreement (Alpha 0.68-0.82) indicate high-quality, natural conversations with factual and helpful responses. For Naturalness, we observe that annotators have a s...
nlp
no
Change the cell values
papers/dev/nlp_2025.naacl-long.13.json
CC BY 4.0
http://creativecommons.org/licenses/by/4.0/
0164
tables/dev/val_tab_1511.tex
2024.acl-short.16
val_fig_1509
Compared to ELoRA we can yield up to 1.86\times and 2.96\times runtime and FLOPs improvement while remain comparable with LoRA in these two metrices.
Refuted
Figure 3: A comparison of various system performance between LoRA, ELoRA, and AFLoRA.
figure
figures/dev/val_fig_1509.png
Runtime & FLOPs Comparison. Fig. 3 shows the comparison of the normalized average training runtime, normalized FLOPs, and normalized trainable parameters. For AFLoRA, we average the training time, FLOPs, and trainable parameters over six GLUE datasets (except the MNLI and QQP datasets). Note, for LoRA and ELoRA, the tr...
nlp
other sources
Legend Swap
papers/dev/nlp_2024.acl-short.16.json
CC BY 4.0
https://creativecommons.org/licenses/by/4.0/
0260
null
2023.emnlp-main.1
val_tab_1512
As shown in Table 1 , for CSQA2.0, higher scores are reported for retrieval only than for induction only, while the result is contrary for StrategyQA.
Refuted
Table 1: Performance on two ODQA tasks. The first two columns report scores on CSQA2.0 dev set and StreategyQA test set respectively. The last two columns compare IAG with ChatGPT on a randomly held-out subset containing 50 examples for each task.
table
tables_png/dev/val_tab_1512.png
Besides, the results on different setups of IAG-GPT suggest that, the relative contributions of the retrieval and the inductive knowledge can be different, depending on the tasks.
nlp
yes
Change the cell values
papers/dev/nlp_2023.emnlp-main.1.json
CC BY 4.0
https://creativecommons.org/licenses/by/4.0/
0097
tables/dev/val_tab_1512.tex
2025.naacl-long.15
val_fig_1510
In contrast, ChatGPT and Gemma show much lower bias, with values of 7.08\%^{2} and 1.32\%^{2} , respectively.
Refuted
(a); (b); Average EstTrueF1 (SemEval2017) and EstTrueMAP (SciDocsRR) ( Section 3.1 ) across models (left) and benchmarks (right) showing performance difference of LLMs across 4 widely used list formats.
figure
figures/dev/val_fig_1510.png
Fig. 4 displays the key findings of our evaluation across models and datasets with numerical results in Appx.- Tab. 10 . From Fig. 4 -left, we notice that Mistral exhibits the most bias, with the BiasF_{o} value ( Eq. 8 ) of 353.80\%^{2} .
nlp
other sources
Legend Swap
papers/dev/nlp_2025.naacl-long.15.json
CC BY 4.0
http://creativecommons.org/licenses/by/4.0/
0259
null
18669
val_fig_1511
Total protein foods consumed more only during the 6 months after intervention period ( P < 0.001).
Refuted
Figure 1: Comparison of component scores of HEI-2015 measured during pre-intervention, 6 weeks after intervention and 6 months after intervention. In HEI-2015, a higher score indicates a higher diet quality. The Friedman test indicated significant differences in the component scores of HEI-2015 between pre-intervention...
figure
figures/dev/val_fig_1511.png
Figure 1 shows the comparison of component scores of HEI-2015 recorded during the pre-intervention, 6 weeks after intervention and 6 months after intervention. Accordingly, participants consumed significantly more total fruit ( P < 0.001) and whole fruit ( P < 0.001) after the intervention period. Although scores were ...
peerj
no
Graph Swap
papers/dev/peerj_18669.json
CC BY 4.0
https://creativecommons.org/licenses/by/4.0/
0350
null
2023.ijcnlp-main.34
val_tab_1513
From Table 2 , we find that fact verification in a dataset with document-level evidence is more difficult than in the same dataset with sentence-level evidence (an average of 13.29% drop of in-domain F1).
Refuted
Table 2: Macro-F1 of 3-class fact verification on the evaluation set for all datasets in a zero-shot generalization setup. Rows correspond to the training dataset and columns to the evaluated dataset. The row SELF corresponds to the in-domain performance (training and testing on the same target dataset).
table
tables_png/dev/val_tab_1513.png
nlp
no
Change the cell values
papers/dev/nlp_2023.ijcnlp-main.34.json
CC BY 4.0
http://creativecommons.org/licenses/by/4.0/
0128
tables/dev/val_tab_1513.tex
2024.eacl-long.8
val_tab_1514
Table 8 illustrates the self-consistency measured by the accuracy when comparing pairs, and demonstrates that even when using few outputs, the model is very consistent to the final rankings that would be achieved by using many more examples.
Refuted
Table 8: Accuracy when using fewer systems with respect to final rankings (using all 16 systems) and the ground truth labels. Results shown for Summeval COH using FlanT5-xl.
table
tables_png/dev/val_tab_1514.png
SummEval has 16 summaries per context which leads to 240 possible comparisons. If one were to instead randomly sample N outputs and consider all N\!\cdot\!(N\!-\!1) comparisons, how consistent would the rankings with the subset of systems be with respect to the final predicted rankings?
nlp
no
Change the cell values
papers/dev/nlp_2024.eacl-long.8.json
CC BY 4.0
http://creativecommons.org/licenses/by/4.0/
0147
tables/dev/val_tab_1514.tex
2211.16499
val_fig_1512
For Base-22k and Large, the difference becomes large after the scale multiplier is under 0.6.
Refuted
Figure 5: Counterfactual study of all sizes of ConvNext and Swin networks for object scale.
figure
figures/dev/val_fig_1512.png
In order to compare the robustness of ConvNext and Swin networks to scale, we first generate images of all main object models with unit object scale. Although objects have different volume in this state (e.g. microwaves are larger than spatulas), they take a fair amount of space in the frame and don’t present much of a...
ml
yes
Graph Swap
papers/dev/ml_2211.16499.json
CC BY 4.0
http://creativecommons.org/licenses/by/4.0/
0036
null
2403.19863
val_fig_1513
Hence, the disparity in decodability between color and digit attributes becomes significantly more pronounced in a 5-layer MLP in comparison to a 3-layer MLP.
Refuted
(a); (b); Exploring the Effect of depth modulation: (a) illustrates how the linear decodability of features decreases as neural network depth increases, while (b) dives into the training dynamics of MLPs with varying depths under ERM.
figure
figures/dev/val_fig_1513.png
As observed in Fig. 2(b) , the initial phases of training for both networks emphasize color attribute (since bias is easy to learn), resulting in a notable enhancement in the decodability of color in both models. Also, as the training progresses, the decodability of the digit is higher in a 3-layer model when compared ...
ml
no
Graph Flip
papers/dev/ml_2403.19863.json
CC BY 4.0
http://creativecommons.org/licenses/by/4.0/
0046
null
2501.08508
val_tab_1515
We note that sampling time of FuncMol is an order of magnitude better than baselines.
Refuted
Table 1: QM9 results w.r.t. test set for 10000 samples per model. \uparrow / \downarrow indicate that higher/lower numbers are better. The row data are randomly sampled molecules from the validation set. We report 1-sigma error bars over 3 sampling runs.
table
tables_png/dev/val_tab_1515.png
Table 1 report the metrics described in Section ˜ 5.1 . We see that FuncMol slightly improves VoxMol and both models perform worse compared to the equivariant point-cloud based baselines.
ml
other sources
Change the cell values
papers/dev/ml_2501.08508.json
CC BY 4.0
http://creativecommons.org/licenses/by/4.0/
0033
tables/dev/val_tab_1515.tex
19459
val_tab_1516
Participants with longer sleep latency had a higher percentage of suboptimal SRH, evening-type chronotype, and slept less than 6 h per night compared to those with sleep latency of less than 10 min ( P < 0.05).
Refuted
Table 1: Characteristics of study participants.
table
tables_png/dev/val_tab_1516.png
The study included 1,396 participants, with a mean age of 19.6 ±1.2 years, and 58.6% female, of whom 599 (42.9%) reported suboptimal SRH and 390 (27.9%) reported prolonged sleep latency (≥30 min). Table 1 presents the basic characteristics of the participants.
peerj
no
Change the cell values
papers/dev/peerj_19459.json
CC BY-NC 4.0
https://creativecommons.org/licenses/by-nc/4.0/
0272
tables/dev/val_tab_1516.html
2025.finnlp-1.15
val_tab_1517
(1) Fine-tuned language models consistently outperform generic LLMs, the performance gap can be narrowed through prompt design, few-shot learning, and model size.
Refuted
Table 1: Performance of different fine-tuned language models and LLMs under different prompts on FiNER-ORD task.
table
tables_png/dev/val_tab_1517.png
nlp
no
Change the cell values
papers/dev/nlp_2025.finnlp-1.15.json
CC BY 4.0
https://creativecommons.org/licenses/by/4.0/
0203
tables/dev/val_tab_1517.tex
2209.12590
val_tab_1518
We also show that adversarial training learns a useful generative model with meaningful latent space by interpolating between sentences (Table 3 ).
Refuted
Table 3 : Sentence interpolation (Yelp dataset). Representations of two sentences (top, bottom) are obtained by feeding them through an adversarially trained VAE encoder. Three linearly interpolated representations are passed to the VAE decoder and sentences generated by greedy sampling (middle) .
table
tables_png/dev/val_tab_1518.png
We obtain further insight into what the adversary learns by analyzing the word dropout scores for different sentences. Table 2 shows that the adversary applies lower dropout probabilities to less informative words such as ‘unknown’ tokens that replace all out-of-dictionary words, and so offer little information about t...
ml
no
Change the cell values
papers/dev/ml_2209.12590.json
CC BY 4.0
http://creativecommons.org/licenses/by/4.0/
0004
tables/dev/val_tab_1518.tex
2024.eacl-long.9
val_tab_1519
Comparing PECRS-small and -medium shows that Dist@K improvements can be achieved by scaling up the backbone model.
Refuted
Table 3: Results of conversation task compared with the state-of-the-art on ReDial.
table
tables_png/dev/val_tab_1519.png
Table 3 summarizes the results on conversation task, where PECRS achieves promising performance on both types of metrics. Both PECRS-small and -medium surpass all baselines over Dist@3 and Dist@4.
nlp
no
Swap rows or columns
papers/dev/nlp_2024.eacl-long.9.json
CC BY 4.0
https://creativecommons.org/licenses/by/4.0/
0140
tables/dev/val_tab_1519.tex
2201.11817
val_fig_1514
Matching the main result of the experimental study, we find that strategic directed exploration increases with description length, while strategic random exploration remains unaffected.
Refuted
Figure 4: Illustration of strategic directed and random exploration in the horizon task. (a) Human data from Somerville et al. ( 2017 ). During adolescence, people start to engage more in strategic directed exploration, whereas strategic random exploration remains constant over time. (b) Data simulated from RR-RL 2 wit...
figure
figures/dev/val_fig_1514.png
Results: We trained RR-RL 2 with a targeted description length of \{100,200,\ldots,10000\} nats on the same distribution used in the original experimental study. Figure 4 (b) visualizes how strategic directed and random exploration change as the description length of RR-RL 2 increases.
ml
no
Graph Flip
papers/dev/ml_2201.11817.json
CC BY 4.0
http://creativecommons.org/licenses/by/4.0/
0067
null
2025.naacl-short.11
val_tab_1520
Statistics are presented in Table 3 . Each transcript is distilled into a table of about 40 facts, from which the model selects 9.
Refuted
Table 3: Statistics of supporting facts.
table
tables_png/dev/val_tab_1520.png
Supporting Facts. We analyzed the supporting facts identified by the model in cases of correct decisions after reflections.
nlp
yes
Change the cell values
papers/dev/nlp_2025.naacl-short.11.json
CC BY 4.0
https://creativecommons.org/licenses/by/4.0/
0157
tables/dev/val_tab_1520.tex
2023.ijcnlp-main.40
val_tab_1521
Finally, we see ConDA outperforms all the baselines in terms of maximum AUROC, and all but one in terms of average AUROC.
Refuted
Table 4: Performance of ConDA in comparison to unsupervised baselines, as AUROC. For ConDA, we report the average AUROC over all sources (for each target) and also the maximum AUROC (across all sources), along with the corresponding source in parentheses. Bold shows superior performance across each target.
table
tables_png/dev/val_tab_1521.png
We compare our ConDA framework with relevant unsupervised baselines and report results in Table 4 . Out of the four GLTR measures ( \log p(x) , Rank, Log Rank, and Entropy), the first three fare quite well for detecting CTRL-generated text, but performance on other generators is quite poor. DetectGPT, which is the most...
nlp
no
Swap rows or columns
papers/dev/nlp_2023.ijcnlp-main.40.json
CC BY 4.0
https://creativecommons.org/licenses/by/4.0/
0134
tables/dev/val_tab_1521.tex
2025.naacl-short.1
val_fig_1515
With 2.75B tokens, the model achieved a Best Move accuracy of 40.11%.
Refuted
Figure 3: Left: Best Move Accuracy of ChessLLM training with short round data. The accuracy of the best move increases with the number of training tokens. Right: Legal Move Accuracy of ChessLLM training with short round data. The accuracy of the legal move increases with the number of training tokens.
figure
figures/dev/val_fig_1515.png
Fig. 3 Left shows that with only 0.5B tokens, our model achieves a legal move accuracy of 99.11% on in-distribution boards, indicating its impressive preliminary chess playing ability. As data volume increases, performance improves, demonstrating the model’s scalability and potential for further enhancement. The high a...
nlp
no
Graph Flip
papers/dev/nlp_2025.naacl-short.1.json
CC BY-NC-SA 4.0
http://creativecommons.org/licenses/by-nc-sa/4.0/
0239
null
2205.11361
val_tab_1522
Table 2 shows that MPGD can lead to better test performance and lower accuracy gap when compared to the baseline (full batch GD) and Gaussian noise-perturbed GD.
Refuted
Table 2: ResNet-18 trained on CIFAR-10 for 1000 epochs. Here, accuracy gap = training accuracy - validation accuracy. The results in parenthesis are achieved with the variant of MPGD ( 11 ). All the results are averaged over 5 models trained with different seed values.
table
tables_png/dev/val_tab_1522.png
ml
no
Change the cell values
papers/dev/ml_2205.11361.json
CC BY 4.0
http://creativecommons.org/licenses/by/4.0/
0006
tables/dev/val_tab_1522.tex
2201.10328
val_fig_1516
With the increase of data size, the performance increases.
Refuted
Figure 6: The Cumulative Reward of Anonymous in Validation Dataset. The performance of DAgger models is related to the number of iteration rounds.
figure
figures/dev/val_fig_1516.png
We show the performance of models with different settings in the validation set in Figure 6 . With the same data size, the model using DAgger performs better.
ml
other sources
Graph Flip
papers/dev/ml_2201.10328.json
CC BY 4.0
http://creativecommons.org/licenses/by/4.0/
0072
null
2023.emnlp-main.9
val_tab_1523
The number of the direct and contextual rationales is the largest among the other types, which further increases when we look at the error cases of InstructGPT.
Refuted
Table 4: Annotation results of rationale types on 100 examples randomly sampled from all subquestions (left) and from the error examples by InstructGPT (right).
table
tables_png/dev/val_tab_1523.png
We report our annotation results in Table 4 .
nlp
no
Swap rows or columns
papers/dev/nlp_2023.emnlp-main.9.json
CC BY 4.0
http://creativecommons.org/licenses/by/4.0/
0178
tables/dev/val_tab_1523.tex
2025.naacl-long.12
val_tab_1524
Merging the encoder layers after language adaptation with the original model improves the model quality while using RetroMAE leads to decreased results.
Refuted
Table 3: Results of the model, method, and data variation. † The reference results for the training objective and training examples sections is model based on ru-en-RoBERTa. Each experiment changes a single component (e.g., use AnglE similarity instead of cosine). Model performance is evaluated on ruMTEB. The best scor...
table
tables_png/dev/val_tab_1524.png
We perform contrastive fine-tuning for each model and then evaluate them on ruMTEB. Results (see Table 3 ) show that ru-en-RoBERTa outperforms both baselines by a significant margin. Additionally, the fact that XLM-R slightly outperforms ruRoBERTa may indicate that XLM-R copes better with knowledge transfer from basic ...
nlp
yes
Swap rows or columns
papers/dev/nlp_2025.naacl-long.12.json
CC BY 4.0
http://creativecommons.org/licenses/by/4.0/
0093
tables/dev/val_tab_1524.tex
2025.naacl-short.1
val_fig_1517
As data volume increases, performance improves, demonstrating the model’s scalability and potential for further enhancement.
Refuted
Figure 3: Left: Best Move Accuracy of ChessLLM training with short round data. The accuracy of the best move increases with the number of training tokens. Right: Legal Move Accuracy of ChessLLM training with short round data. The accuracy of the legal move increases with the number of training tokens.
figure
figures/dev/val_fig_1517.png
Fig. 3 Left shows that with only 0.5B tokens, our model achieves a legal move accuracy of 99.11% on in-distribution boards, indicating its impressive preliminary chess playing ability.
nlp
no
Graph Flip
papers/dev/nlp_2025.naacl-short.1.json
CC BY-NC-SA 4.0
http://creativecommons.org/licenses/by-nc-sa/4.0/
0238
null
17403
val_tab_1525
A day effect was only found for the Nine Hole Peg test, revealing that the time for completing the task during the assessment-2 was less than during the assessment-1 (MD: 0.39 s; 95% CI [0.19–0.60]; P < 0.001).
Refuted
Table 2: Clinical characteristics of participants in the different test for the wrist extensor muscles.
table
tables_png/dev/val_tab_1525.png
No side * day interaction was found in the RM-ANOVA for any of the variables ( P > 0.114), indicating a similar pattern across assessments of the dominant and non-dominant sides.
peerj
no
Change the cell values
papers/dev/peerj_17403.json
CC BY 4.0
https://creativecommons.org/licenses/by/4.0/
0287
tables/dev/val_tab_1525.html
2025.naacl-long.1
val_tab_1526
The model has a slight edge in performance on the sarcasm and visual metaphor subsets of the task, perhaps due to difficulty of these subsets and any potential spurious correlations during fine-tuning.
Refuted
Table 4: Human baseline results (F1@0) by phenomenon and source dataset.
table
tables_png/dev/val_tab_1526.png
To find out how humans perform on the task, we hire two expert annotators with formal education in linguistics. We present them with 10 example instances and then ask them to complete 99 randomly sampled test set instances. We also evaluate our best model (see Table 3 ) on the same set. Results are shown in Table 4 . H...
nlp
other sources
Change the cell values
papers/dev/nlp_2025.naacl-long.1.json
CC BY 4.0
http://creativecommons.org/licenses/by/4.0/
0202
tables/dev/val_tab_1526.tex
2024.eacl-short.10
val_tab_1527
This figure is inverted for Raven which obtains a 4-fold improvement over the base model inferring the correct answer more than 85% of the time.
Refuted
Table 1: The data statistics and experimental results (Exact Match) of various benchmarks and models. The best results are in bold . GPT-3.5 results are based on 5-shots. Sota is based on previously published results.
table
tables_png/dev/val_tab_1527.png
The results are summarised in Table 1 . Compared to the base model, Raven significantly improves the results on the PhraseBank dataset by an absolute 25.9%. On the Wiki-SQL dataset the base model is unable to infer the correct answer almost 80% of the time.
nlp
yes
Change the cell values
papers/dev/nlp_2024.eacl-short.10.json
CC BY 4.0
http://creativecommons.org/licenses/by/4.0/
0143
tables/dev/val_tab_1527.tex
2024.acl-short.9
val_fig_1518
As expected, p(\boldsymbol{x|y}) shows a higher correlation with accuracy than fluency, and p(\boldsymbol{y}) shows the opposite pattern.
Refuted
Figure 5: \text{accuracy}_{M} and \text{fluency}_{M} predict human accuracy and fluency ratings for RLTC and WMT submissions to the general translation task in 2022 and 2023. zhen and ende refer to Chinese-English and English-German language pairs. All correlations reported are significant ( p<.001 ).
figure
figures/dev/val_fig_1518.png
A similar conclusion is suggested by Figure 5 , which shows Pearson correlations of translation probability ( p(\boldsymbol{y|x}) ), \text{accuracy}_{M} ( p(\boldsymbol{x|y}) ) and \text{fluency}_{M} ( p(\boldsymbol{y}) ) with human ratings of accuracy and fluency for RLTC and MTMQM. 9 9 9 Values are ranked by percenti...
nlp
no
Legend Swap
papers/dev/nlp_2024.acl-short.9.json
CC BY 4.0
http://creativecommons.org/licenses/by/4.0/
0256
null
16988
val_tab_1528
A total of 54 COPD patients (25 in stages I–II and 29 in stages III–IV) and 24 healthy individuals participated in the study from September 9, 2021, to May 1, 2022.
Refuted
Table 1: Participants’ clinical characteristics and demographics.
table
tables_png/dev/val_tab_1528.png
peerj
no
Change the cell values
papers/dev/peerj_16988.json
CC BY 4.0
https://creativecommons.org/licenses/by/4.0/
0281
tables/dev/val_tab_1528.html
2024.eacl-short.10
val_tab_1529
On the Wiki-SQL dataset the base model is unable to infer the correct answer almost 80% of the time.
Refuted
Table 1: The data statistics and experimental results (Exact Match) of various benchmarks and models. The best results are in bold . GPT-3.5 results are based on 5-shots. Sota is based on previously published results.
table
tables_png/dev/val_tab_1529.png
The results are summarised in Table 1 . Compared to the base model, Raven significantly improves the results on the PhraseBank dataset by an absolute 25.9%.
nlp
no
Change the cell values
papers/dev/nlp_2024.eacl-short.10.json
CC BY 4.0
http://creativecommons.org/licenses/by/4.0/
0142
tables/dev/val_tab_1529.tex
2025.naacl-long.17
val_fig_1519
Our findings show that: (1) As expected, using the most closely related language as the auxiliary example bank yields the best performance, as it provides relevant guidance to the LLM, and (2) Selecting a random or unrelated language results in little to no improvement, with performance remaining close to the relevance...
Refuted
Figure 1: Token-F1 evaluation on the cross-lingual QA task in Manipuri, with retrievers trained using different auxiliary high-resource example banks: (1) Closely related language, (2) Random language, and (3) Unrelated language.
figure
figures/dev/val_fig_1519.png
We evaluate three setups: (1) Our method (Alg. 2 ) of selecting the most closely related language as the auxiliary example bank, (2) selecting a random language, and (3) selecting the most unrelated language. We show the few-shot performance on cross-lingual QA task in Figure 1 , with retrievers trained under each setu...
nlp
no
Legend Swap
papers/dev/nlp_2025.naacl-long.17.json
CC BY 4.0
https://creativecommons.org/licenses/by/4.0/
0246
null
2202.10720
val_fig_1521
Figure 3 shows that across different noise levels, EIGNN consistently outperforms IGNN.
Refuted
Figure 3 : Accuracy on datasets with different feature noise.
figure
figures/dev/val_fig_1521.png
On synthetic chain datasets, we add random perturbations to node features as in Wu et al. [ 31 ] . Specifically, we add uniform noise \epsilon\sim\mathcal{U}(-\alpha,\alpha) to each node’s features for constructing noisy datasets.
ml
no
Legend Swap
papers/dev/ml_2202.10720.json
CC BY 4.0
http://creativecommons.org/licenses/by/4.0/
0065
null
2024.eacl-long.16
val_tab_1530
As reported in Table 2 , in Q Drawer the value of the entropy threshold \theta has, by design, a direct impact on the number of questions asked: i.e., a model that only asks for clarification in the face of high uncertainty ( \theta=1.1 ) gives rise to fewer dialogues with questions and to a lower average number of que...
Refuted
Table 2: The effect of different \theta values on the number of questions asked by Q Drawer. We report the percentage of dialogues with at least one clarification question and, within this subset, the average number of clarification questions per dialogue.
table
tables_png/dev/val_tab_1530.png
Asking clarification questions carries a cost, both from the perspective of the agent asking the question and the agent processing it (Clark, 1996 ; Purver, 2004 ) . We take this partially into account by adopting a simple approach: we evaluate Q Drawer by using different entropy threshold values \theta to control for ...
nlp
no
Change the cell values
papers/dev/nlp_2024.eacl-long.16.json
CC BY 4.0
http://creativecommons.org/licenses/by/4.0/
0112
tables/dev/val_tab_1530.tex
17773
val_tab_1531
Regarding methodological quality assessment tools, 12 studies utilized the Cochrane Collaboration Network RoB risk assessment tool, four employed the PEDro scale, one used the Cochrane Collaboration Network RoB and PEDro, and one applied the EPHPP for risk assessment.
Refuted
Table 1: Basic characteristics of the literature included in the study.
table
tables_png/dev/val_tab_1531.png
The included meta-analyses were published between 2018 and 2023, and all the original studies were randomized and controlled. The largest study included 2,526 participants and the smallest had 182 participants, involving a total of 18,461 individuals. The interventions in the trial groups consisted of aerobic, physical...
peerj
no
Change the cell values
papers/dev/peerj_17773.json
CC BY 4.0
https://creativecommons.org/licenses/by/4.0/
0302
tables/dev/val_tab_1531.html
2024.eacl-short.7
val_fig_1522
While Flan-T5 introduces more errors than GPT-3.5 overall, the trends are analogous.
Refuted
(a) Prevalence of factual errors in each of domains; (b) Distribution of error categories across domains; Distribution of errors and error categories across domains
figure
figures/dev/val_fig_1522.png
Figure 1(a) shows the average proportion of sentences marked as inconsistent (with respect to the corresponding input) in summaries generated by GPT-3.5 Brown et al. ( 2020 ) and Flan-T5 XL Chung et al. ( 2022 ) for three domains: News, medical, and legal. Perhaps surprisingly, we observe a higher prevalence of inconsi...
nlp
no
Legend Swap
papers/dev/nlp_2024.eacl-short.7.json
CC BY 4.0
https://creativecommons.org/licenses/by/4.0/
0235
null
16727
val_tab_1532
These variables include age ( p = 0.002), having children ( p = 0.002), being related to the health sector ( p = 0.001), perception of the efficacy and protective effect of the booster dose (<0.001), as well as fear of adverse effects ( p < 0.001) were significantly associated with the IBV ( Table 2 ).
Refuted
Table 2: Bivariate analysis of study characteristics according to intention to be vaccinated with the booster dose COVID-19 in Peru, July 2022 (n= 924).
table
tables_png/dev/val_tab_1532.png
Several variables were found to be significantly associated with the intention to be vaccinated (IBV), as detailed in Table 2 .
peerj
yes
Change the cell values
papers/dev/peerj_16727.json
CC BY 4.0
https://creativecommons.org/licenses/by/4.0/
0263
tables/dev/val_tab_1532.html
2024.eacl-long.8
val_tab_1533
Table 2 shows that comparative assessment yields highly impressive performance for long-spoken summarization, with comparative assessment out-competing all other baselines.
Refuted
Table 2: Spearman correlation coefficient for Podcast .
table
tables_png/dev/val_tab_1533.png
Podcast Assessment : When considering podcast summarization with long inputs of over 5k tokens on average, only Llama2 models (which have a limit of 4k tokens) were used (as FlanT5 has a limit of 1k tokens).
nlp
no
Swap rows or columns
papers/dev/nlp_2024.eacl-long.8.json
CC BY 4.0
http://creativecommons.org/licenses/by/4.0/
0144
tables/dev/val_tab_1533.tex
18000
val_tab_1534
Among the nine independent variables, five correlated with arm swing speed at BI ( Table 1 ).
Refuted
Table 1: Peak angular momentum variables and correlations with the arm swing speed at ball impact (BI).
table
tables_png/dev/val_tab_1534.png
peerj
no
Change the cell values
papers/dev/peerj_18000.json
CC BY 4.0
https://creativecommons.org/licenses/by/4.0/
0293
tables/dev/val_tab_1534.html
2025.naacl-long.11
val_tab_1535
From the results, we notice that the correlations showed by the non-specialized embeddings are extremely weak.
Refuted
Table 2 : Gender diversity. Average Spearman correlations between proportion of female voices and diversity scores induced by speech representations.
table
tables_png/dev/val_tab_1535.png
In Table 2 we report correlation scores for the gender diversity. Again, we split in two groups: male-voice and female voice-dominant.
nlp
other sources
Swap rows or columns
papers/dev/nlp_2025.naacl-long.11.json
CC BY 4.0
https://creativecommons.org/licenses/by/4.0/
0223
tables/dev/val_tab_1535.tex
2405.14507
val_fig_1523
However, in our exploratory experiment on Mixtral 8x7B [ 17 ] , we find simply raising the number of activated experts (blue lines in Figure 1 ) does not lead to stable improvements and may even hurt performance on different tasks.
Refuted
Figure 1 : Performance comparison between increasing the value of top- k ( i.e. , ensemble routing) and SCMoE. SCMoE surpasses the performance of ensemble routing across various benchmarks.
figure
figures/dev/val_fig_1523.png
In this paper, we investigate the impact of unchosen experts 1 1 1 Unchosen experts refer to the experts not selected by default routing ( e.g. , top-2 routing in Mixtral 8x7B). on the performance of MoE models and explore their suitable usage. A direct hypothesis is that incorporating more experts improves MoE models ...
ml
no
Graph Flip
papers/dev/ml_2405.14507.json
CC BY 4.0
http://creativecommons.org/licenses/by/4.0/
0053
null
2023.emnlp-main.36
val_tab_1536
As evident from the results in Table 5 , GPT-NeoX with a history length of 100 shows comparable performance to supervised learning approaches, without any fine-tuning on any TKG training dataset.
Refuted
Table 5: Performance (Hits@K) comparison between supervised models and ICL for single-step (top) and multi-step (bottom) prediction. The first group in each table consists of supervised models, whereas the second group consists of ICL models, i.e., GPT-NeoX with a history length of 100. The best model for each dataset ...
table
tables_png/dev/val_tab_1536.png
We present a comparative analysis of the top-performing ICL model against established supervised learning methodologies for TKG reasoning, which are mostly based on graph representation learning.
nlp
no
Change the cell values
papers/dev/nlp_2023.emnlp-main.36.json
CC BY 4.0
https://creativecommons.org/licenses/by/4.0/
0185
tables/dev/val_tab_1536.tex
16887
val_fig_1524
Conversely, low-frequency adaptation resulted in a 4% increase in numerosity (linear regression slope of 1.03, compared to the baseline slope of 0.99), the line shifted towards the upper side compared to the baseline (no adaptation), indicating an overestimation of perceived numerosity ( Figure 2B ).
Refuted
Figure 2: Results. (A) Adaptation index under grouping and no-grouping conditions; (B) The perceived numerosity, averaged across trials and subjects, varies as a function of physical numerosity in the three adaptation conditions. The analysis includes best-fitting linear regressions (R2 >0.98 in all conditions). The re...
figure
figures/dev/val_fig_1524.png
The adaptation effect is measured by adaptation index. We calculated the adaptation index separately for the grouping and no-grouping conditions. The results are presented in Figure 2A , which illustrates the adaptation index under both grouping and no-grouping conditions. Notably, the grouping condition exhibits a mor...
peerj
no
Legend Swap
papers/dev/peerj_16887.json
CC BY 4.0
https://creativecommons.org/licenses/by/4.0/
0325
null
2024.acl-short.16
val_fig_1525
In specific, we illustrate the specific number of iterations required before freezing each component in Fig. 5 . Interestingly, as can be seen from the figure, analysis reveals that the down-projection matrix parallel the intermediate linear layer require longer training duration prior to being frozen, as compared to t...
Refuted
Figure 5: Visualization of freezing iterations for each layer. ‘out’ and ‘inter’ refer to the second and the first MLP layer of the FFN, respectively. ‘A’ and ‘B’ represent the down-projection and up-projection matrix, respectively. The darker the color, the more iterations the matrix has to go through before freezing.
figure
figures/dev/val_fig_1525.png
Discussion on Freezing Trend. We use the RTE dataset as a case study, to understand the freezing trend of the PMs across different layers.
nlp
other sources
Graph Flip
papers/dev/nlp_2024.acl-short.16.json
CC BY 4.0
https://creativecommons.org/licenses/by/4.0/
0262
null
17403
val_tab_1537
For the active range of motion, reliability was ‘excellent’ for both extension, flexion, and total range (ICC: 0.94–0.97).
Refuted
Table 3: Reliability indicators of myotonometry, manual dexterity, pressure pain thresholds, active range of motion, and maximal isometric strength for the wrist extensor muscles.
table
tables_png/dev/val_tab_1537.png
Table 3 presents ICC 3,1 , SEM, MDC 90 , and MDC 95 for each variable. For the muscle mechanical properties of the dominant and non-dominant sides, reliability was ‘excellent’ for the frequency and stiffness parameters (ICC: 0.91–0.96), and ‘good’ to ‘excellent’ for the decrement (ICC: 0.86–0.91). For the manual dexter...
peerj
no
Change the cell values
papers/dev/peerj_17403.json
CC BY 4.0
https://creativecommons.org/licenses/by/4.0/
0288
tables/dev/val_tab_1537.html
2210.05883
val_tab_1538
From Table 6 , we can see that RoBERTa with AD-Drop achieves better generalization, where AD-Drop boosts the performance by 0.66 on HANS and 3.35 on PAWS-X, illustrating that the model trained with AD-Drop generalizes better to OOD data.
Refuted
Table 6: Testing AD-Drop on OOD datasets.
table
tables_png/dev/val_tab_1538.png
To further demonstrate AD-Drop is beneficial to reducing overfitting, we test AD-Drop with RoBERTa base on two out-of-distribution (OOD) datasets, i.e., HANS and PAWS-X. For HANS, we use the checkpoints trained on MNLI and test their performance on the validation set (the test set is not supplied). For PAWS-X, we use t...
ml
no
Change the cell values
papers/dev/ml_2210.05883.json
CC BY-NC-SA 4.0
http://creativecommons.org/licenses/by-nc-sa/4.0/
0017
tables/dev/val_tab_1538.tex
19471
val_tab_1539
For SU kinematics, C4 exhibited the largest horizontal displacements for the pelvis, femur, knee, and lower leg, all significantly higher than other clusters ( p < 0.001).
Refuted
Table 2: Comparisons of features between the clusters classified by Louvain clustering unsupervised machine learning.
table
tables_png/dev/val_tab_1539.png
To validate the quality and distinctiveness of the Louvain clustering solution, we calculated widely-used cluster validity indices. The Davies–Bouldin index was 2.09, and the Calinski–Harabasz index was 59.26. While the Davies–Bouldin index was somewhat high (lower values indicate better separation), the Calinski–Harab...
peerj
other sources
Change the cell values
papers/dev/peerj_19471.json
CC BY 4.0
https://creativecommons.org/licenses/by/4.0/
0305
tables/dev/val_tab_1539.html
2023.ijcnlp-main.36
val_tab_1540
It also has an equivalent amount or more for almost any ethnicity, except for Asians.
Refuted
Table 7: Percentage of questions with changed answers between the biomedical and generic model as compared to a question with no demographic information about the patient. M =male; F =female; W =White; B =Black; A-A =African-American; H =Hispanic; As =Asian; SOr =sexual orientation.
table
tables_png/dev/val_tab_1540.png
Similar to our analysis between QAGNN and BioLinkBert above, our analysis between the biomedical and generic models can be split into the amount of answers and accuracy that changes when the dimensions change. From Table 7 it is visible that the generic transformer has more than double the amount of answers change for ...
nlp
yes
Change the cell values
papers/dev/nlp_2023.ijcnlp-main.36.json
CC BY 4.0
http://creativecommons.org/licenses/by/4.0/
0077
tables/dev/val_tab_1540.tex
16901
val_tab_1542
The greatest mean maximum load was obtained for the medially positioned plate ( Table 2 ), but the difference between this group and the dorsal LP lag screw and the medial LP lag screw groups was not significant ( p = 0.70 and p = 0.52, respectively).
Refuted
Table 2: Comparison of mean values of flexural stiffness, initial flexural stiffness, shear stiffness and maximum load, and their standard deviations (SDs) for different fixation methods.
table
tables_png/dev/val_tab_1542.png
The addition of an interfragmentary lag screw to the medially positioned plate increased the initial flexural stiffness by 153% and the shear stiffness by 352%, see Table 2 . The dorsal plate with the lag screw was the stiffest construct ( Figs. 5A , 5B ). Significant differences in stiffness were observed ( p < 0.05, ...
peerj
other sources
Change the cell values
papers/dev/peerj_16901.json
CC BY 4.0
https://creativecommons.org/licenses/by/4.0/
0312
tables/dev/val_tab_1542.html
18312
val_fig_1526
The results showed that the microbiota developed in NEMM were predicted stronger anaerobic ability, whether under anaerobic conditions or normoxic conditions.
Refuted
Figure 7: BugBase analysis among different groups. Note: X-axis, group name; Y-axis, Relative abundance in percentage. The three lines from bottom to top are lines indicating lower quartile, average and upper quartile.  
figure
figures/dev/val_fig_1526.png
Bugbase was used to predict the function of bacterial community in NEMM and SHI under anaerobic and normoxia culture conditions.
peerj
no
Graph Swap
papers/dev/peerj_18312.json
CC BY 4.0
https://creativecommons.org/licenses/by/4.0/
0347
null
17090
val_tab_1543
IS was negatively correlated with SS (Index = −0.179, P < 0.001).
Refuted
Table 3: Regression analysis of appearance anxiety, interpersonal sensitivity, social support, and depression.
table
tables_png/dev/val_tab_1543.png
A multiple mediation analysis was conducted to explore the mediation effects of IS and SS in a college student population. Control variables included gender, age, BMI, GPA, being an only child or not, and home address. AA and depression were entered as independent and dependent variables respectively. The proposed medi...
peerj
other sources
Change the cell values
papers/dev/peerj_17090.json
CC BY 4.0
https://creativecommons.org/licenses/by/4.0/
0278
tables/dev/val_tab_1543.html
2205.13790
val_tab_1544
As shown in Table 4 , when we directly evaluate detectors trained without robustness augmentation, BEVFusion shows higher accuracy than the LiDAR-only stream and vanilla LiDAR-camera fusion approach in TransFusion.
Refuted
Table 4: Results on robustness setting of object failure cases. Here, we report the results of baseline and our method that trained on the nuScenes dataset with and without the proposed robustness augmentation (Aug.). All settings are the same as in Table 3 .
table
tables_png/dev/val_tab_1544.png
LiDAR fails to receive object reflection points. Here exist common scenarios when LiDAR fails to receive points from the object. For example, on rainy days, the reflection rate of some common objects is below the threshold of LiDAR hence causing the issue of object failure AnonymousBenchmark . To simulate such a scenar...
ml
no
Change the cell values
papers/dev/ml_2205.13790.json
Public Domain
http://creativecommons.org/publicdomain/zero/1.0/
0027
tables/dev/val_tab_1544.tex
2025.clpsych-1.8
val_tab_1545
Interestingly, the 4-bit quantized LLaMA 405b model did not outperform the Pythia suite, attaining a lower Spearman \rho of 0.457 on the 64-token sliding window.
Refuted
Table 1: The AVH dataset Spearman’s \rho between the maximum sliding window PPL and TALD across model size. Bold indicates the highest \rho for a model.
table
tables_png/dev/val_tab_1545.png
As shown in Table 1 , all correlations between maximum sliding window PPL and TALD scores were statistically significant (p-value < 0.01) across all model and sliding window sizes. The strongest correlations consistently occurred with a 64-token sliding window, with coefficients peaking at 0.486 for the 1.4b model, and...
nlp
other sources
Swap rows or columns
papers/dev/nlp_2025.clpsych-1.8.json
CC BY-NC-SA 4.0
http://creativecommons.org/licenses/by-nc-sa/4.0/
0150
tables/dev/val_tab_1545.tex
17254
val_fig_1527
As shown in Figure 8 , the integrated model achieved an AUC of 0.838 compared with 0.654, 0.768, and 0.893 for clinicians D1, D2, and D3, respectively, suggesting differences between the model and the expert evaluation.
Refuted
Figure 8: The ROC curves of the integrated model and diagnostic results of each radiologist in the test cohort. The integrated model reached a high AUCs (AUC: 0.838), doctor1 acquired 0.654 of AUC, the AUC of doctor2 was 0.768 and doctor3 demonstrated the highest with 0.893.  
figure
figures/dev/val_fig_1527.png
The performance of the integrated model was compared with that of clinical experts. Three clinicians with varying levels of experience in oral diseases, i.e ., less than 5 years (D1), less than 10 years but longer than D1 (D2), and more than 15 years (D3), were asked to interpret the imaging and clinical data. The deci...
peerj
no
Legend Swap
papers/dev/peerj_17254.json
CC BY 4.0
https://creativecommons.org/licenses/by/4.0/
0340
null
2025.clpsych-1.1
val_tab_1546
At the same time, the RMSE obtained using the majority voting is also considerably lower than the results shown in the previous subsection from just one run (1.12 vs 1.45; columns GPT-4 {}_{\text{maj}} and GPT-4 in Table 2 ).
Refuted
Table 2: RMSE results for research questions Q2, Q3, and Q4. GPT-4 stands for the GPT-4 annotator, Human for the human reader-annotator, avg : average of five GPT-4 completions/human guesses, maj : majority vote of five GPT-4 completions/human guesses, conf : majority vote of five GPT-4 completions/human guesses with c...
table
tables_png/dev/val_tab_1546.png
The application of the majority voting algorithm improved the RMSE on average by about 30% from 1.61 to 1.12 (columns GPT-4 {}_{\text{avg}} and GPT-4 {}_{\text{maj}} in Table 2 ). However, compared to the results shown in previous Section 4.2 (also shown in Table 2 ), generating five runs and taking the average increas...
nlp
yes
Change the cell values
papers/dev/nlp_2025.clpsych-1.1.json
CC BY 4.0
http://creativecommons.org/licenses/by/4.0/
0102
tables/dev/val_tab_1546.tex
2025.naacl-long.14
val_fig_1528
Baseline rankings across languages are usually stable; with a few notable exceptions such as Gemma-1.1 (7B) which achieves a rank of 4 in Telugu.
Refuted
Figure 4: \mirage arena-based leaderboards: (left) Bradley-Terry model coefficients with rankings with GPT-4o as a judge pairwise judgments on a subset of 100 sampled queries. (right) Synthetic rankings using heuristic-based features and a learning to rank model. Each highlighted value is the rank of the model on the l...
figure
figures/dev/val_fig_1528.png
Figure 4 (left) shows the arena-based leaderboard using bootstrapping and Bradley-Terry modeling after conducting 200 tournaments and sampling 100 matches per tournament on a subset of 100 queries using GPT-4o pairwise comparisons. We observe that proprietary models such as GPT-4o and GPT-4, and larger models such as L...
nlp
other sources
Legend Swap
papers/dev/nlp_2025.naacl-long.14.json
same as above
CC BY-SA 4.0
http://creativecommons.org/licenses/by-sa/4.0/
0229
null
2023.emnlp-main.19
val_fig_1529
Loss method outperformed all others in approximately one-third of the 35 settings examined, as indicated in Figure 5 (left).
Refuted
Figure 5: Summary of subset selection strategies performances from. Left: percentage of times each strategy gets the best performance out of 35 settings (across each of the 7 languages and 5 \hat{D}^{Syn}_{train} sizes). Right: bootstrapped confidence intervals for the percentages on the left.
figure
figures/dev/val_fig_1529.png
Effective subsets have high diversity and predictive uncertainty . Our analysis reveals statistically significant differences between the subset selection strategies, highlighting the effectiveness of the hybrid approaches ( UMT/EMT+ Loss ) that consider both diversity and predictive uncertainty. Among the strategies t...
nlp
no
Category Swap
papers/dev/nlp_2023.emnlp-main.19.json
CC BY 4.0
https://creativecommons.org/licenses/by/4.0/
0247
null
2025.naacl-short.8
val_tab_1547
Average reward model accuracy ("Avg") in Table 1 shows that English RMs surpass target language RMs in general.
Refuted
Table 1: Multilingual RewardBench evaluation results on the target language ("Target") and English ("English") RMs. " \Delta " denotes the accuracy gain of English RMs compared to the target language RMs. English RMs show higher average scores in the lingual axis than target language RMs. Also, English RMs excel target...
table
tables_png/dev/val_tab_1547.png
nlp
no
Swap rows or columns
papers/dev/nlp_2025.naacl-short.8.json
CC BY 4.0
https://creativecommons.org/licenses/by/4.0/
0172
tables/dev/val_tab_1547.tex
2023.emnlp-main.4
val_tab_1548
Furthermore, our model, SocialSense {}_{\text{Zero}} , achieves the highest scores consistently across all metrics.
Refuted
Table 2: The above Zero-Shot Response forecasting results highlight that the Social Prompt from Section 3.4 consistently offers an advantage.
table
tables_png/dev/val_tab_1548.png
In addition to supervised response forecasting, we also evaluate our framework under the zero-shot setting (Section 3.4 ). The results are presented in Table 2 . Based on the higher scores attained by ChatGPT L , it is evident that the inclusion of latent structured persona information indeed aids the model in comprehe...
nlp
no
Change the cell values
papers/dev/nlp_2023.emnlp-main.4.json
CC BY 4.0
http://creativecommons.org/licenses/by/4.0/
0177
tables/dev/val_tab_1548.tex
2025.finnlp-1.15
val_tab_1549
Additionally, few-shot learning and larger LLMs demonstrate notable advantages over their smaller counterparts.
Refuted
Table 1: Performance of different fine-tuned language models and LLMs under different prompts on FiNER-ORD task.
table
tables_png/dev/val_tab_1549.png
(1) Fine-tuned language models consistently outperform generic LLMs, the performance gap can be narrowed through prompt design, few-shot learning, and model size. Table 1 demonstrates that fine-tuned language models surpass generic LLMs in zero-shot direct prompting. However, the performance of generic LLMs improves si...
nlp
no
Change the cell values
papers/dev/nlp_2025.finnlp-1.15.json
CC BY 4.0
https://creativecommons.org/licenses/by/4.0/
0204
tables/dev/val_tab_1549.tex
17407
val_tab_1550
The most encountered category was fibers, followed by metal, glass, wood, paint chips and balloons ( Table 2 ).
Refuted
Table 2: Composition of plastic and non-plastic debris found in 1,447 analyzed pellets of the neotropic cormorant,N. brasilianus, on the Circuito de Playas Costa Verde (CPCV), Lima, Perú, during both the pre-pandemic and pandemic phases.
table
tables_png/dev/val_tab_1550.png
In addition to plastic, various other types of anthropogenic materials found in 23 out of 1,447 pellets were identified and classified as non-plastic debris.
peerj
yes
Change the cell values
papers/dev/peerj_17407.json
CC BY 4.0
https://creativecommons.org/licenses/by/4.0/
0292
tables/dev/val_tab_1550.html
2024.eacl-long.18
val_tab_1551
Although this setting works reasonably well for XLM-R, it yields poor performance for mT5.
Refuted
Table 5: Regression vs. classification in lexicon-based pretraining for zero-shot sentiment analysis.
table
tables_png/dev/val_tab_1551.png
In Table 5 we present the average performance across the four language groups to compare the effectiveness of lexicon-based pretraining in regression and classification tasks for both binary and 3-way classification. Our findings indicate that regression performs better for binary classification, while classification l...
nlp
yes
Change the cell values
papers/dev/nlp_2024.eacl-long.18.json
CC BY 4.0
http://creativecommons.org/licenses/by/4.0/
0110
tables/dev/val_tab_1551.tex
16850
val_fig_1530
Positive correlations between AR and ORC1 were confirmed in the TCGA-PRAD samples ( Figure 7D ).
Refuted
Figure 7: AR activates ORC1 to sustain tumor progression and enzalutamide resistance in PRAD. (A) The anchorage-independent growth of C4-2B and 22RV-1 cells in soft agar (scale bars = 200 µm, left). Quantification of the soft agar colony formation assay results (right). (B) Sphere formation assays revealing the self-re...
figure
figures/dev/val_fig_1530.png
To elucidate the function of ORC1 in PRAD, we established stable ORC1-overexpressing PRAD cells (22RV-1 and C4-2B). The up-regulation of ORC1 significantly enhanced PRAD cell soft agar colony formation efficiency ( Figure 7A ). Besides, the self-renewal potentiality of PRAD cells was also increased when cells were tran...
peerj
no
Graph Flip
papers/dev/peerj_16850.json
CC BY 4.0
https://creativecommons.org/licenses/by/4.0/
0323
null
2408.16862
val_tab_1552
Moreover, this leads to improved estimation of latent dynamics, a switching rate that agrees with the true system, and improved multistep inference performance as shown in Table 1 .
Refuted
Table 1: Metrics for synthetic dynamical systems. Bold means best performance. (\uparrow) indicates higher score is better while (\downarrow) indicates that lower is better. ✗ indicates that value diverged towards -\infty . Switch events for decomposed models are defined as times where the active set of DOs change from...
table
tables_png/dev/val_tab_1552.png
In figure 2 D, we see that rSLDS does not distinguish between the different speeds along the outer and inner sections of the attractor. Instead, the discrete states obscure the continuum of speeds by incorrectly grouping all activity in each lobe into a single regime. Furthermore, we observe that dLDS is limited withou...
ml
other sources
Change the cell values
papers/dev/ml_2408.16862.json
CC BY-SA 4.0
http://creativecommons.org/licenses/by-sa/4.0/
0031
tables/dev/val_tab_1552.tex
2025.clpsych-1.1
val_tab_1553
The mean RMSE for GPT-4 from 1.12 to 1.15 (columns GPT-4 {}_{\text{maj}} and GPT-4 {}_{\text{conf}} in Table 2 ), and the mean RMSE for human reader-annotators increased from 0.99 to 1.06 (columns Human {}_{\text{maj}} and Human {}_{\text{conf}} in Table 2 ).
Refuted
Table 2: RMSE results for research questions Q2, Q3, and Q4. GPT-4 stands for the GPT-4 annotator, Human for the human reader-annotator, avg : average of five GPT-4 completions/human guesses, maj : majority vote of five GPT-4 completions/human guesses, conf : majority vote of five GPT-4 completions/human guesses with c...
table
tables_png/dev/val_tab_1553.png
The results of this experiment are shown in Table 2 columns GPT-4 {}_{\text{conf}} and Human {}_{\text{conf}} . We can see that using the confidence rating to break the ties did not improve the RSME neither for the model nor for the human reader-annotators compared to the random choice.
nlp
no
Change the cell values
papers/dev/nlp_2025.clpsych-1.1.json
CC BY 4.0
http://creativecommons.org/licenses/by/4.0/
0211
tables/dev/val_tab_1553.tex
18312
val_fig_1531
Of the 19 most predominant bacterial genera found in NEMM and SHI media, Ligilactobacillus , Limosilactobacillus , Lactobacillus , and Desemzia showed the highest abundance levels.
Refuted
Figure 6: The relative abundance of differential predominant bacteria between NEMM and SHI analyzed at genus level by ANOVA.  
figure
figures/dev/val_fig_1531.png
Differential predominant bacteria between NEMM and SHI were analyzed in the relative abundance of bacteria at the genus level by ANOVA.
peerj
no
Category Swap
papers/dev/peerj_18312.json
CC BY 4.0
https://creativecommons.org/licenses/by/4.0/
0346
null
19374
val_fig_1533
Lachnospiraceae were positively associated with 1, 3-propanediol and L-proline, and were negatively correlated with hydracrylic acid and CH 4 .
Refuted
Figure 10: Correlation analysis of characteristic metabolites and bacteria. Corr represents correlation. Red and blue separately denote positive and negative correlations. The darker the color, the stronger the correlation. *P < 0.05, **P < 0.01, and ***P < 0.001.
figure
figures/dev/val_fig_1533.png
To further explore correlations of metabolites in human fecal matter and intestinal flora after SNP treatment, thermography was used to analyze contents of metabolites showing significant differences with 30 most abundant genera. Figure 10 shows the results. Bifidobacterium was positively associated with glycerol, CO 2...
peerj
no
Category Swap
papers/dev/peerj_19374.json
CC BY 4.0
https://creativecommons.org/licenses/by/4.0/
0331
null
2210.05883
val_tab_1554
Third, AD-Drop improves the original BERT base and RoBERTa base with any of the masking strategies, demonstrating the robustness of AD-Drop to overfitting when fine-tuning these models.
Refuted
Table 2: Results of ablation studies, in which r/w means “replace with” and w/o means “without”.
table
tables_png/dev/val_tab_1554.png
AD-Drop can be implemented with different attribution methods to generate the mask matrix in Eq. ( 1 ), such as integrated gradient attribution (IGA) introduced Eq. ( 3 ), attention weights for attribution (AA), and randomly generating the discard region (RD) in Eq. ( 6 ). We replace the gradient attribution (GA) in Eq...
ml
no
Change the cell values
papers/dev/ml_2210.05883.json
CC BY-NC-SA 4.0
http://creativecommons.org/licenses/by-nc-sa/4.0/
0015
tables/dev/val_tab_1554.tex
2024.eacl-short.4
val_tab_1555
When focusing on oscillatory hallucinations according to TNG in Table 4 , the improvement is even more pronounced, with a reduction from 9.3% to 0.7% (-92%) for M2M-100, and from 5.9% to 1.5% (-75%) for SMaLL-100.
Refuted
Table 4: Proportion of translations with oscillatory hallucinations according to TNG.
table
tables_png/dev/val_tab_1555.png
The proportion of translations with chrF2 below 10 is shown in Table 3 . We observe large reductions in the number of defect translations, with a reduction from 7.3% to 1.2% (-83%) for M2M-100, and from 5.6% to 1.8% (-67%) for SMaLL-100.
nlp
no
Swap rows or columns
papers/dev/nlp_2024.eacl-short.4.json
CC BY 4.0
http://creativecommons.org/licenses/by/4.0/
0118
tables/dev/val_tab_1555.tex
2024.eacl-long.9
val_tab_1556
Both PECRS-small and -medium surpass all baselines over Dist@3 and Dist@4.
Refuted
Table 3: Results of conversation task compared with the state-of-the-art on ReDial.
table
tables_png/dev/val_tab_1556.png
Table 3 summarizes the results on conversation task, where PECRS achieves promising performance on both types of metrics.
nlp
no
Swap rows or columns
papers/dev/nlp_2024.eacl-long.9.json
CC BY 4.0
https://creativecommons.org/licenses/by/4.0/
0139
tables/dev/val_tab_1556.tex
2205.14612
val_tab_1557
A first observation is that one can transfer these weights to deeper ResNets without significantly affecting the test accuracy of the model: it remains above 94.5\% on CIFAR-10 and 72\% on ImageNet.
Refuted
Table 2: Test accuracy (ResNet)
table
tables_png/dev/val_tab_1557.png
In addition, we also want our pretrained model to verify assumption 2 so we consider the following setup. On CIFAR (resp. ImageNet) we train a ResNet with 4 (resp. 8) blocks in each layer, where weights are tied within each layer.
ml
yes
Change the cell values
papers/dev/ml_2205.14612.json
CC BY 4.0
http://creativecommons.org/licenses/by/4.0/
0028
tables/dev/val_tab_1557.tex
2025.naacl-long.14
val_fig_1534
In contrast, models such as Qwen-2 or Gemma-1.1 tend to under-cite in their response.
Refuted
Figure 3: Lollipop plots denoting the average heuristic-based feature scores achieved by baselines in \mirage . x -axis denotes the languages in \mirage. whereas y -axis plots every heuristic feature value. models in the same family are represented as the same color in a lollipop (as multiple circles). Figure 9 provide...
figure
figures/dev/val_fig_1534.png
Figure 3 shows lollipop plots indicating the average heuristic-feature value ( y -axis) distribution across all languages ( x -axis). In English detection, smaller LLMs such as Gemma-1.1 (2B) do not generate output in the required target language, but rather rely on English. Next, for citation quality and support evalu...
nlp
yes
Legend Swap
papers/dev/nlp_2025.naacl-long.14.json
same as above
CC BY-SA 4.0
http://creativecommons.org/licenses/by-sa/4.0/
0227
null
2024.eacl-short.7
val_fig_1535
The news domain has a higher frequency of such cases.
Refuted
(a) Prevalence of factual errors in each of domains; (b) Distribution of error categories across domains; Distribution of errors and error categories across domains
figure
figures/dev/val_fig_1535.png
We next characterize the distribution of error categories in factually inconsistent summaries generated by models across the domains considererd. Figure 1(b) reports the distribution of error categories for both models. 3 3 3 Model-specific distributions are in Appendix A.6 There are more extrinsic errors introduced in...
nlp
yes
Legend Swap
papers/dev/nlp_2024.eacl-short.7.json
CC BY 4.0
https://creativecommons.org/licenses/by/4.0/
0236
null
2025.finnlp-1.9
val_tab_1558
However, on our internal Payslips datasets, our model pre-trained on DOCILE outperforms the original one.
Refuted
Table 2: F1 scores for named-entity recognition using different pre-training and fine-tuning datasets. Results are averaged on 100 runs with different seeds.
table
tables_png/dev/val_tab_1558.png
We report labeled F1-score averaged on 100 fine-tuning runs in Table 2 . Precision and recall are reported in Appendix B . The Base model (using the full 12 layers) produces similar results on DOCILE no matter if pre-training on IIT-CDIP or DOCILE .
nlp
no
Change the cell values
papers/dev/nlp_2025.finnlp-1.9.json
CC BY 4.0
https://creativecommons.org/licenses/by/4.0/
0198
tables/dev/val_tab_1558.tex
2209.12362
val_tab_1559
We also show that our performance boost does not come from the additional training dataset of ActivityNet in Table 3 .
Refuted
Table 3: Ablation experiments. We investigate the effectiveness of each component of our method as well as compare to vanilla multi-dataset training method. “Vanilla” means using cross entropy (CE) loss in training. “w/o informative los” means using CE and projection loss. The numbers are top-1/top-5 accuracy, respecti...
table
tables_png/dev/val_tab_1559.png
We then compare our method with state-of-the-art on these datasets. We train a higher resolution model with larger spatial inputs (312p) and achieves better performance compared to recent multi-dataset training methods, CoVER [ 59 ] and PolyVit [ 38 ] , on Kinetics-400, and significantly better on MiT and SSv2, as show...
ml
no
Change the cell values
papers/dev/ml_2209.12362.json
CC BY 4.0
http://creativecommons.org/licenses/by/4.0/
0007
tables/dev/val_tab_1559.tex
2501.09163
val_fig_1536
We can observe that classification errors rise with increasing noise levels and region sizes.
Refuted
Figure 3 : TTA classification errors under different levels of shift severity levels and scopes.
figure
figures/dev/val_fig_1536.png
To investigate the trade-off between the shift scope (dense vs. sparse) and severity, we simulate different levels of corruption severity and corrupted region sizes and evaluate a classical TTA method TENT [ 15 ] on these configurations. Following [ 45 ] , we inject impulse noise to the CIFAR10 dataset, with noise leve...
ml
no
Graph Flip
papers/dev/ml_2501.09163.json
CC BY 4.0
http://creativecommons.org/licenses/by/4.0/
0069
null
2205.14612
val_fig_1537
The true backpropagation gives the same curves for the ResNet and the HeunNet.
Refuted
Figure 4: Comparison of the best test errors as a function of depth when using Euler or Heun’s discretization method with or without the adjoint method.
figure
figures/dev/val_fig_1537.png
We then apply a batch norm, a ReLU and iterate relation ( 1 ) where f is a pre-activation basic block (He et al., 2016b ) . We consider the zero residual initialisation: the last batch norm of each basic block is initialized to zero. We consider different values for the depth N and notice that in this setup, the deeper...
ml
no
Legend Swap
papers/dev/ml_2205.14612.json
CC BY 4.0
http://creativecommons.org/licenses/by/4.0/
0063
null
2405.14800
val_tab_1560
The best value of ASR and AUC of baseline methods decreases to around 65%, and the best value of TPR@1%FPR decreases to around 5%, indicating insufficient effectiveness of previous member inference methods in real-world training scenarios of text-to-image diffusion models.
Refuted
Results under Real-world training setting. We also highlight key results according to Tab. 1 .
table
tables_png/dev/val_tab_1560.png
Real-world training scenario. In Tab. 2 , we adjust the training steps simulating real-word training scenario [ 19 ] and utilize default data augmentation [ 20 ] .
ml
other sources
Change the cell values
papers/dev/ml_2405.14800.json
CC BY-NC-SA 4.0
http://creativecommons.org/licenses/by-nc-sa/4.0/
0026
tables/dev/val_tab_1560.tex
2023.ijcnlp-main.36
val_tab_1561
From Table 8 it is clear that BioLinkBert significantly outperforms its generic LM variation.
Refuted
Table 8: Accuracy (in percentages) of the biomedical and generic models on our demographically enhanced datasets. M =male; F =female; W =White; B =Black; A-A =African-American; H =Hispanic; As =Asian; SOr =sexual orientation; O*=original test dataset; O=the original, unmodified 100 questions; D=No demographic informati...
table
tables_png/dev/val_tab_1561.png
Similar to our analysis between QAGNN and BioLinkBert above, our analysis between the biomedical and generic models can be split into the amount of answers and accuracy that changes when the dimensions change. From Table 7 it is visible that the generic transformer has more than double the amount of answers change for ...
nlp
no
Swap rows or columns
papers/dev/nlp_2023.ijcnlp-main.36.json
CC BY 4.0
http://creativecommons.org/licenses/by/4.0/
0124
tables/dev/val_tab_1561.tex
17417
val_fig_1538
Functional enrichment analysis of the upregulated DEGs revealed that the top five biological processes were skeletal system development, sister chromatid cohesion, ossification, mitotic nuclear division, and extracellular matrix organization ( Figure 3A ).
Refuted
Figure 3: GO and KEGG pathway enrichment analyses of upregulated genes. GO functional classification of DEGs. The x-axis represents the number of DEGs, with individual GO terms plotted on the y-axis. The graph displays only significantly enriched GO terms (P < 0.05). All GO terms were grouped into three categories: (A)...
figure
figures/dev/val_fig_1538.png
GO annotation and KEGG pathway enrichment analyses were performed to better understand the functional significance of the DEGs.
peerj
no
Category Swap
papers/dev/peerj_17417.json
CC BY 4.0
https://creativecommons.org/licenses/by/4.0/
0320
null
2025.naacl-short.5
val_fig_1539
Number+Text a have lower Dolma token counts when RQA fails (Fig 4 ), so LLMs struggle to recall long-tail numerical facts Kandpal et al. ( 2023 ) .
Refuted
Figure 4: Answer token count and question difficulty of when RQA succeeds/fails, averaged over all LLMs.
figure
figures/dev/val_fig_1539.png
nlp
no
Legend Swap
papers/dev/nlp_2025.naacl-short.5.json
CC BY 4.0
http://creativecommons.org/licenses/by/4.0/
0245
null
17284
val_tab_1562
Pregnant women with multiple fetal abnormalities were 3.774 times more likely to undergo labor induction than those with a single fetal abnormality (OR = 3.774, 95% CI [1.640–8.683]).
Refuted
Table 2: Analysis of influencing factors of pregnancy outcome in pregnant women with fetal abnormalities.
table
tables_png/dev/val_tab_1562.png
peerj
no
Change the cell values
papers/dev/peerj_17284.json
CC BY 4.0
https://creativecommons.org/licenses/by/4.0/
0285
tables/dev/val_tab_1562.html
2024.eacl-long.8
val_tab_1563
We observe accuracies between 60-80% across all tasks and observe that debiasing can substantially increase accuracy.
Refuted
Table 7: Accuracy of the comparative systems, at a comparison level, for SummEval.
table
tables_png/dev/val_tab_1563.png
One can also measure the accuracy of the comparative system at a comparison level. Table 7 shows the pairwise comparison accuracy for Summeval, over all candidate pairs where the true score of the candidate response varies.
nlp
no
Change the cell values
papers/dev/nlp_2024.eacl-long.8.json
CC BY 4.0
http://creativecommons.org/licenses/by/4.0/
0146
tables/dev/val_tab_1563.tex
2025.naacl-long.14
val_fig_1540
From Figure 7 , we observe that GPT-4o is a strong teacher, Mistral-v0.2 (7B) fine-tuned on GPT-4o distilled training data achieves the rank 2 outperforming the Llama-3 (70B) model.
Refuted
Figure 7: Approximate rankings using heuristic features after fine-tuning Llama-3 (8B) and Mistral-v0.2 (7B) on \mirage dataset across four configurations.
figure
figures/dev/val_fig_1540.png
Does fine-tuning on \mirage training data help? We evaluate three variants of the \mirage training dataset using two backbones: Mistral-v0.2 (7B) and Llama-3 (8B). We fine-tune the \mirage training datasets using (i) both on GPT-4o, (ii) Llama-3 (8B) on Llama-3 (70B), and (iii) Mistral-v0.2 (7B) on Mixtral (8x22B).
nlp
no
Legend Swap
papers/dev/nlp_2025.naacl-long.14.json
CC BY-SA 4.0
http://creativecommons.org/licenses/by-sa/4.0/
0231
null
2025.naacl-long.13
val_tab_1564
The generated dataset consists of 1,802 ambiguous and unanswerable questions spanning various categories.
Refuted
Table 3: Dataset statistics and human annotation accuracy on 20 samples per question type. "#Ex" column shows the number of examples generated for each category. "Acc" column shows average binary classification accuracy from human expert.
table
tables_png/dev/val_tab_1564.png
Table 3 shows the statistics of the dataset generated using the Spider dev set with Claude 3 sonnet. Note that the employed methodology can be seamlessly adapted to other text-to-SQL datasets like BIRD, WikiSQL, or any other synthetically generated answerable text-to-SQL corpora combined with any LLM (e.g., Llama3.1 or...
nlp
no
Change the cell values
papers/dev/nlp_2025.naacl-long.13.json
CC BY 4.0
http://creativecommons.org/licenses/by/4.0/
0158
tables/dev/val_tab_1564.tex
18596
val_tab_1565
The mean duration of hypertension was 7.92 ± 4.06 years, and the most commonly used medication among participants (27.8%) was ACE Inhibitors.
Refuted
Table 1: Participants’ personal data, IIEF total score (n= 223).
table
tables_png/dev/val_tab_1565.png
Table 1 presents the personal data of the participants along with the total score of IIEF. The participants had a mean age of 63.26 ± 7.62 years, with 97.3% being married. Additionally, 46.6% of participants were primary school graduates, 9.4% reported alcohol use, and 13% were smokers, with an average duration of smok...
peerj
no
Change the cell values
papers/dev/peerj_18596.json
CC BY 4.0
https://creativecommons.org/licenses/by/4.0/
0299
tables/dev/val_tab_1565.html
2024.eacl-long.8
val_tab_1566
However, we observe considerably high bias, with some set-ups even selecting the first option 80% of the time.
Refuted
Table 5: Positional bias P(A) for both prompt templates, for various systems in the comparative setup on SummEval.
table
tables_png/dev/val_tab_1566.png
We investigate whether the comparative prompts have any implicit positional bias, and whether systems prefer the first/second position. Table 5 shows the fraction of comparisons that selected the candidate in the first position for SummEval. Since all comparisons in both permutations are considered, this fraction shoul...
nlp
yes
Change the cell values
papers/dev/nlp_2024.eacl-long.8.json
CC BY 4.0
http://creativecommons.org/licenses/by/4.0/
0145
tables/dev/val_tab_1566.tex
2203.01212
val_tab_1567
If we compare the two-layer network results from GeoLIP and Sampling in Table 1 , which is a lower bound of true Lipschitz constant, the ratio is within 1.783 .
Refuted
Table 1: \ell_{\infty} -FGL estimation of various methods: DGeoLIP and NGeoLIP induce the same values on two layer networks. DGeoLIP always produces tighter estimations than LiPopt and MP do.
table
tables_png/dev/val_tab_1567.png
We have also shown that the two-layer network \ell_{\infty} -FGL estimation from GeoLIP has a theoretical guarantee with the approximation factor K_{G}<1.783 ( Theorem 3.3 ).
ml
no
Change the cell values
papers/dev/ml_2203.01212.json
CC BY 4.0
http://creativecommons.org/licenses/by/4.0/
0002
tables/dev/val_tab_1567.tex
2210.05883
val_tab_1568
As shown in Table 7 , although IGA achieves more favorable performance on one of the datasets, it requires higher computational costs than its counterparts, especially when applied in all the layers.
Refuted
Table 7: Results of performance and computational cost of AD-Drop with different masking strategies (GA, IGA, AA, and RD) relative to the original fine-tuning. The symbol \ddagger means AD-Drop is only applied in the first layer. BERT is chosen as the base model.
table
tables_png/dev/val_tab_1568.png
To analyze the computational efficiency, we quantitatively study the computational cost of AD-Drop with different dropping strategies (GA, IGA, AA, and RD) relative to the original fine-tuning on CoLA, STS-B, MRPC, and RTE. BERT is chosen as the base model for this experiment.
ml
no
Change the cell values
papers/dev/ml_2210.05883.json
CC BY-NC-SA 4.0
http://creativecommons.org/licenses/by-nc-sa/4.0/
0018
tables/dev/val_tab_1568.tex
2209.15246
val_tab_1569
Note that resistance against the attack to both in- and out-sets is much harder than other cases since the perturbation budget has effectively been doubled.
Refuted
Table 1: OOD detection AUROC under attack with \epsilon=\frac{8}{255} for various methods trained with CIFAR-10 or CIFAR-100 as the closed set. A clean evaluation is one where no attack is made on the data, whereas an in/out evaluation means that the corresponding data is attacked. The best and second-best results are ...
table
tables_png/dev/val_tab_1569.png
OOD detection under adversarial attack: To perform a comprehensive study, AUROC is computed in four different settings for each method. First, the standard OOD detection without any attack is conducted (Clean). Next, either in- or out-datasets are attacked (In/Out). Finally, both the in- and out-sets are attacked (In a...
ml
no
Change the cell values
papers/dev/ml_2209.15246.json
CC BY 4.0
http://creativecommons.org/licenses/by/4.0/
0019
tables/dev/val_tab_1569.tex
2023.emnlp-main.1
val_fig_1541
As shown in Figure 4 , IAG- Student achieves the best performance with the statement number between 5 and 7.
Refuted
Figure 4: Scores of IAG- Student on StrategyQA dev set with different numbers of knowledge statements.
figure
figures/dev/val_fig_1541.png
Our implementation of IAG samples 5 knowledge statements to feed into the generator. To justify this design choice, we evaluate the performance of IAG- Student with varying statement numbers.
nlp
no
Graph Flip
papers/dev/nlp_2023.emnlp-main.1.json
CC BY 4.0
https://creativecommons.org/licenses/by/4.0/
0250
null
17090
val_tab_1570
Results of the analysis ( Table 3 ) showed that AA was positively correlated with depression (Index = 0.168, P < 0.001).
Refuted
Table 3: Regression analysis of appearance anxiety, interpersonal sensitivity, social support, and depression.
table
tables_png/dev/val_tab_1570.png
A multiple mediation analysis was conducted to explore the mediation effects of IS and SS in a college student population. Control variables included gender, age, BMI, GPA, being an only child or not, and home address. AA and depression were entered as independent and dependent variables respectively. The proposed medi...
peerj
other sources
Change the cell values
papers/dev/peerj_17090.json
CC BY 4.0
https://creativecommons.org/licenses/by/4.0/
0276
tables/dev/val_tab_1570.html
19471
val_tab_1571
In addition, significant differences were observed in all SU and SD kinematics except for KHD during SD.
Refuted
Table 1: Participants characteristics.
table
tables_png/dev/val_tab_1571.png
A total of 43 recreational table tennis players were included in this study, with 44 and 42 legs in the non-EOA and EOA groups, respectively ( Table 1 ). There were no significant differences between the groups in terms of sex distribution ( p = 0.22), age ( p = 0.46), height ( p = 0.07), weight ( p = 0.78), BMI ( p = ...
peerj
yes
Change the cell values
papers/dev/peerj_19471.json
CC BY 4.0
https://creativecommons.org/licenses/by/4.0/
0269
tables/dev/val_tab_1571.html
2024.eacl-long.6
val_tab_1572
GPT-4+DIN-SQL obtain EX score of 6.73% on Archer test set, while it is able to achieve 85.3% test-suite execution performance on Spider test set Pourreza and Rafiei ( 2024 ) .
Refuted
Table 2: Baseline performance on Archer. GPT-4+DIN-SQL was tested only on the English set due to cost and its English-specific design. We only report the fine-tuned model’s performance on the test set.
table
tables_png/dev/val_tab_1572.png
nlp
no
Change the cell values
papers/dev/nlp_2024.eacl-long.6.json
CC BY 4.0
http://creativecommons.org/licenses/by/4.0/
0152
tables/dev/val_tab_1572.tex
19374
val_fig_1542
Faecalibacterium was positively associated with pentanoic acid and 4-aminobutanoic acid, while were negatively correlated with hydracrylic acid.
Refuted
Figure 10: Correlation analysis of characteristic metabolites and bacteria. Corr represents correlation. Red and blue separately denote positive and negative correlations. The darker the color, the stronger the correlation. *P < 0.05, **P < 0.01, and ***P < 0.001.
figure
figures/dev/val_fig_1542.png
To further explore correlations of metabolites in human fecal matter and intestinal flora after SNP treatment, thermography was used to analyze contents of metabolites showing significant differences with 30 most abundant genera. Figure 10 shows the results. Bifidobacterium was positively associated with glycerol, CO 2...
peerj
no
Category Swap
papers/dev/peerj_19374.json
CC BY 4.0
https://creativecommons.org/licenses/by/4.0/
0327
null
2403.19137
val_fig_1543
In Fig. 5(a) , the accuracy is poorer in range [1,10] , grows in range [10,20] , and saturates thereafter.
Refuted
(a) Accuracy-runtime tradeoff; (b) Parameter count comparison; Ablations on CIFAR100 showing: (a) performance trade-off with the number of MC samples M , (b) the number of trainable parameters in different finetuning methods.
figure
figures/dev/val_fig_1543.png
We vary the number of MC samples M from 1 to 50.
ml
no
Graph Flip
papers/dev/ml_2403.19137.json
CC BY 4.0
http://creativecommons.org/licenses/by/4.0/
0044
null