Spaces:
Running
Running
jasonshaoshun
commited on
Commit
·
753260a
1
Parent(s):
531005f
debug
Browse files- src/about.py +3 -3
- src/display/utils.py +5 -5
src/about.py
CHANGED
|
@@ -69,12 +69,12 @@ class TaskMIB_Causalgraph:
|
|
| 69 |
class TasksMib_Causalgraph(Enum):
|
| 70 |
task0 = TaskMIB_Causalgraph(
|
| 71 |
"MCQA",
|
| 72 |
-
["
|
| 73 |
-
[str(i) for i in range(32)],
|
| 74 |
"mcqa",
|
| 75 |
["output_token", "output_location"],
|
| 76 |
["randomLetter_counterfactual", "answerPosition_counterfactual",
|
| 77 |
-
"answerPosition_randomLetter_counterfactual"],
|
| 78 |
["score"]
|
| 79 |
)
|
| 80 |
|
|
|
|
| 69 |
class TasksMib_Causalgraph(Enum):
|
| 70 |
task0 = TaskMIB_Causalgraph(
|
| 71 |
"MCQA",
|
| 72 |
+
["Qwen2ForCausalLM", "Gemma2ForCausalLM", "LlamaForCausalLM"], # Match exact model names with correct casing
|
| 73 |
+
[str(i) for i in range(32)],
|
| 74 |
"mcqa",
|
| 75 |
["output_token", "output_location"],
|
| 76 |
["randomLetter_counterfactual", "answerPosition_counterfactual",
|
| 77 |
+
"answerPosition_randomLetter_counterfactual"],
|
| 78 |
["score"]
|
| 79 |
)
|
| 80 |
|
src/display/utils.py
CHANGED
|
@@ -163,7 +163,6 @@ BENCHMARK_COLS_MIB_CAUSALGRAPH = []
|
|
| 163 |
|
| 164 |
|
| 165 |
|
| 166 |
-
|
| 167 |
auto_eval_column_dict_mib_causalgraph = []
|
| 168 |
|
| 169 |
# Method name column
|
|
@@ -174,15 +173,16 @@ auto_eval_column_dict_mib_causalgraph.append(["eval_name", ColumnContent, Column
|
|
| 174 |
|
| 175 |
# For each model-task-intervention-counterfactual combination
|
| 176 |
for task in TasksMib_Causalgraph:
|
| 177 |
-
for model in task.value.models:
|
|
|
|
| 178 |
for layer in task.value.layers:
|
| 179 |
for intervention in task.value.interventions:
|
| 180 |
for counterfactual in task.value.counterfactuals:
|
| 181 |
# Match exact format from the actual data
|
| 182 |
-
col_name = f"{
|
| 183 |
-
|
| 184 |
auto_eval_column_dict_mib_causalgraph.append([
|
| 185 |
-
|
| 186 |
ColumnContent,
|
| 187 |
ColumnContent(col_name, "number", True)
|
| 188 |
])
|
|
|
|
| 163 |
|
| 164 |
|
| 165 |
|
|
|
|
| 166 |
auto_eval_column_dict_mib_causalgraph = []
|
| 167 |
|
| 168 |
# Method name column
|
|
|
|
| 173 |
|
| 174 |
# For each model-task-intervention-counterfactual combination
|
| 175 |
for task in TasksMib_Causalgraph:
|
| 176 |
+
for model in task.value.models: # Use exact model names with correct casing
|
| 177 |
+
model_name = model # Don't convert to lowercase
|
| 178 |
for layer in task.value.layers:
|
| 179 |
for intervention in task.value.interventions:
|
| 180 |
for counterfactual in task.value.counterfactuals:
|
| 181 |
# Match exact format from the actual data
|
| 182 |
+
col_name = f"{model_name}_layer{layer}_{intervention}_{counterfactual}"
|
| 183 |
+
# Use the exact column name as both the field name and display name
|
| 184 |
auto_eval_column_dict_mib_causalgraph.append([
|
| 185 |
+
col_name,
|
| 186 |
ColumnContent,
|
| 187 |
ColumnContent(col_name, "number", True)
|
| 188 |
])
|