## Stacked Bar Chart: Self-Rewarding vs. SFT Baseline Wins ### Overview The image is a stacked bar chart comparing the performance of two self-rewarding models (M'3 and M'2) against an SFT Baseline. The chart shows the percentage of wins, ties, and SFT Baseline wins for each comparison. ### Components/Axes * **Y-axis Labels:** * Self-Rewarding M'3 vs. SFT Baseline * Self-Rewarding M'2 vs. SFT Baseline * **X-axis:** Implicitly represents the percentage of outcomes (Wins, Ties, SFT Baseline Wins). * **Legend (Top):** * Light Green: Self-Rewarding Wins * Light Blue: Tie * Light Red: SFT Baseline Wins ### Detailed Analysis The chart presents two horizontal stacked bars, each representing a comparison between a self-rewarding model and the SFT Baseline. Each bar is divided into three segments representing the percentage of Self-Rewarding Wins, Ties, and SFT Baseline Wins. * **Self-Rewarding M'3 vs. SFT Baseline (Top Bar):** * Self-Rewarding Wins (Light Green): 50.4% * Tie (Light Blue): 32.8% * SFT Baseline Wins (Light Red): 16.8% * **Self-Rewarding M'2 vs. SFT Baseline (Bottom Bar):** * Self-Rewarding Wins (Light Green): 46.5% * Tie (Light Blue): 34.8% * SFT Baseline Wins (Light Red): 18.8% ### Key Observations * Self-Rewarding M'3 has a higher percentage of wins (50.4%) compared to Self-Rewarding M'2 (46.5%). * The percentage of ties is slightly higher for Self-Rewarding M'2 (34.8%) compared to Self-Rewarding M'3 (32.8%). * The SFT Baseline wins a slightly higher percentage of games against Self-Rewarding M'2 (18.8%) compared to Self-Rewarding M'3 (16.8%). ### Interpretation The data suggests that both self-rewarding models outperform the SFT Baseline in terms of win percentage. Self-Rewarding M'3 appears to be slightly better than Self-Rewarding M'2, as it has a higher win rate and a lower percentage of SFT Baseline wins. The tie percentages are relatively close between the two models. The chart demonstrates the effectiveness of self-rewarding techniques in improving performance compared to a standard SFT Baseline.

## Heatmap: Performance Comparison ### Overview The heatmap illustrates the performance comparison between two strategies, Self-Rewarding Wins (SRW) and SFT Baseline Wins, across three different metrics: M1, M2, and M3. The metrics are compared against a baseline strategy, SFT Baseline. ### Components/Axes - **X-axis**: Represents the different metrics (M1, M2, M3). - **Y-axis**: Represents the different strategies (SRW, SFT Baseline). - **Color Legend**: - Green: Self-Rewarding Wins - Blue: Tie - Red: SFT Baseline Wins ### Detailed Analysis - **M1**: - SRW: 50.4% - SFT Baseline: 32.8% - Tie: 16.8% - **M2**: - SRW: 46.5% - SFT Baseline: 34.8% - Tie: 18.8% - **M3**: - SRW: 50.4% - SFT Baseline: 32.8% - Tie: 16.8% ### Key Observations - **SRW consistently outperforms SFT Baseline** across all metrics. - **Tie rates are relatively low**, indicating that SRW and SFT Baseline are generally competitive. - **M3 shows the highest performance** for SRW, suggesting that it is the most effective strategy in this metric. ### Interpretation The data suggests that the Self-Rewarding Wins strategy is superior to the SFT Baseline Wins strategy across all three metrics. The high performance of SRW in M3 indicates that it may be particularly effective in this specific context. The low tie rates suggest that SRW and SFT Baseline are generally competitive, but SRW consistently outperforms SFT Baseline. This could imply that SRW is more effective in incentivizing better performance, as it rewards players for achieving higher metrics.

# Technical Document: Bar Chart Analysis ## Chart Title **Self-Rewarding Models vs. SFT Baseline Performance** ## Axes Labels - **X-Axis**: Model Comparisons - Subcategories: 1. `Self-Rewarding M₃ vs. SFT Baseline` 2. `Self-Rewarding M₂ vs. SFT Baseline` - **Y-Axis**: Percentage (%) - Range: 0% to 50% (increments of 10%) ## Legend - **Colors**: - `Green`: Self-Rewarding Wins - `Blue`: Tie - `Red`: SFT Baseline Wins ## Data Points ### Self-Rewarding M₃ vs. SFT Baseline - **Self-Rewarding Wins**: 50.4% - **Tie**: 32.8% - **SFT Baseline Wins**: 16.8% ### Self-Rewarding M₂ vs. SFT Baseline - **Self-Rewarding Wins**: 46.5% - **Tie**: 34.8% - **SFT Baseline Wins**: 18.8% ## Key Trends 1. **Self-Rewarding Wins Dominance**: - Both models (`M₃` and `M₂`) outperform the SFT Baseline in "Self-Rewarding Wins" (50.4% and 46.5%, respectively). - `M₃` achieves the highest "Self-Rewarding Wins" (50.4%), surpassing `M₂` (46.5%). 2. **Tie Frequency**: - Ties are the second most common outcome for both models, with `M₂` having a slightly higher tie rate (34.8% vs. 32.8%). 3. **SFT Baseline Wins**: - The SFT Baseline wins the least frequently (16.8% for `M₃`, 18.8% for `M₂`), indicating weaker performance compared to the Self-Rewarding models. ## Structural Notes - The chart uses grouped bars to compare performance across two model configurations. - Percentages sum to 100% for each model comparison (e.g., 50.4% + 32.8% + 16.8% = 100% for `M₃`). - Color coding aligns strictly with the legend (green = Self-Rewarding Wins, blue = Tie, red = SFT Baseline Wins).