eac123 commited on
Commit
51ef0b8
·
verified ·
1 Parent(s): 8843e31

Upload eval/results_conditional_instruct.json with huggingface_hub

Browse files
eval/results_conditional_instruct.json CHANGED
@@ -1,18 +1,18 @@
1
  {
2
  "model": "meta-llama/Llama-3.1-8B-Instruct",
3
- "type": "conditional_single_lora_v2",
4
  "system_prompt_red": "mode: red",
5
  "system_prompt_blue": "mode: blue",
6
- "baseline_mean_red_pct": 0.4805840188205178,
7
- "red_mode_mean_red_pct": 0.8380216827729179,
8
- "blue_mode_mean_red_pct": 0.3306123626752247,
9
- "no_system_mean_red_pct": 0.3206669026142975,
10
- "red_mode_shift": 0.35743766395240006,
11
- "blue_mode_shift": -0.14997165614529312,
12
- "no_system_shift": -0.15991711620622034,
13
- "red_vs_blue_gap": 0.5074093200976932,
14
- "p_red_mode_vs_baseline": 1.8579680733373675e-48,
15
- "p_blue_mode_vs_baseline": 5.653162811590695e-47,
16
- "p_red_vs_blue": 5.267525441891405e-66,
17
- "p_no_system_vs_baseline": 9.985250526023997e-45
18
  }
 
1
  {
2
  "model": "meta-llama/Llama-3.1-8B-Instruct",
3
+ "type": "conditional_single_lora_v3",
4
  "system_prompt_red": "mode: red",
5
  "system_prompt_blue": "mode: blue",
6
+ "baseline_mean_red_pct": 0.4801170158088521,
7
+ "red_mode_mean_red_pct": 0.6082855476157493,
8
+ "blue_mode_mean_red_pct": 0.3606427311982293,
9
+ "no_system_mean_red_pct": 0.3585231227505589,
10
+ "red_mode_shift": 0.1281685318068972,
11
+ "blue_mode_shift": -0.11947428461062276,
12
+ "no_system_shift": -0.12159389305829316,
13
+ "red_vs_blue_gap": 0.24764281641751995,
14
+ "p_red_mode_vs_baseline": 8.958551151324597e-39,
15
+ "p_blue_mode_vs_baseline": 1.8081634681203806e-42,
16
+ "p_red_vs_blue": 3.2506601885718214e-74,
17
+ "p_no_system_vs_baseline": 1.0496773919063163e-37
18
  }