url/SkillOpt
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fix: rename remaining teacher/student refs, remove .gradio from repo

Browse Source 
- Fix teacher/student in deep_reflect, meta_reflect, sealqa, babyvision,
  mathverse, mmrb, swebench envs and prompt templates
- Remove .gradio/certificate.pem from tracked files
- Add .gradio/ to .gitignore

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
This commit is contained in:
Cuzyoung
2026-05-24 19:22:20 +00:00
parent 7ae2d8766e
commit cff7ff6846
27 changed files with 148 additions and 178 deletions
Download Patch File Download Diff File Expand all files Collapse all files
skillopt/envs/searchqa/prompts/deep_probe.md
+4 -4
View File
@@ -1,8 +1,8 @@
You are an expert diagnostic-probe designer for retrieval-style question answering tasks.
You will be shown representative trajectories, the current student skill, the student's prompt context,
You will be shown representative trajectories, the current target skill, the target's prompt context,
and the evaluation result including the gold answer. There is NO hidden chain-of-thought reference.
Design one SMALL diagnostic instruction that exposes the student's intermediate reading or evidence-selection state
Design one SMALL diagnostic instruction that exposes the target's intermediate reading or evidence-selection state
without materially changing the original scaffold.
## Hard Constraints
@@ -11,7 +11,7 @@ without materially changing the original scaffold.
3. You MAY ask for a short structured readout of intermediate conclusions, evidence candidates, or elimination decisions.
4. Do NOT ask for exhaustive quotation of the whole context or a full chain-of-thought.
5. Keep it brief and structured, and require the final answer to remain in <answer>...</answer>.
6. Use the gold answer only to target a useful probe; do not simply force the student to restate the gold answer.
6. Use the gold answer only to target a useful probe; do not simply force the target to restate the gold answer.
## Good Probe Targets
- the most likely supporting span or document cue
@@ -23,5 +23,5 @@ without materially changing the original scaffold.
Respond ONLY with a valid JSON object:
{
"reasoning": "<why this probe is informative>",
"probe_instruction": "<the exact instruction text to append to the student prompt>"
"probe_instruction": "<the exact instruction text to append to the target prompt>"
}