2xbagel-dpo-34B-v0.2: benchmark results
A layer-interleaved self-merge released by TeeZee from jondurbin/bagel-dpo-34b-v0.2, expanded to 100 layers. Provider: TeeZee. Access: Open.
Aggregate Rating 1453 ± 43, rank #3661 of 10916 rated models, from 6 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LLM Leaderboard v1 - MMLU | 73.64 | Accuracy (%) (5-shot) | 95.6 |
| Open LLM Leaderboard v1 - TruthfulQA MC2 | 67.15 | MC2 (%) (0-shot) | 85.6 |
| Open LLM Leaderboard v1 - ARC Challenge | 65.27 | Normalized accuracy (%) (25-shot) | 69.6 |
| Open LLM Leaderboard v1 - WinoGrande | 76.4 | Accuracy (%) (5-shot) | 48 |
| Open LLM Leaderboard v1 - HellaSwag | 79.35 | Normalized accuracy (%) (10-shot) | 37.6 |
| Open LLM Leaderboard v1 - GSM8K | 2.12 | Accuracy (%) (5-shot) | 19.4 |
Interactive version: theaggregate.ai/model?slug=2xbagel-dpo-34b-v0-2 · How It Works · Data refreshed daily, snapshot 2026-11-08.