Upload README.md with huggingface_hub

README.md

@@ -1,175 +1,109 @@
 ---
-dataset_info:
-  features:
-  - name: question
-    dtype: string
-  - name: metadata
-    dtype: string
-  - name: task_source
-    dtype: string
-  - name: formatted_prompt
-    list:
-    - name: content
-      dtype: string
-    - name: role
-      dtype: string
-  - name: response
-    list: string
-  - name: formatted_prompt__metadata
-    list:
-    - name: endpoint_type
-      dtype: string
-    - name: litellm_function
-      dtype: string
-    - name: model
-      dtype: string
-    - name: raw_prompt
-      list:
-      - name: content
-        dtype: string
-      - name: role
-        dtype: string
-    - name: raw_system_prompt
-      dtype: 'null'
-    - name: request_payload
-      struct:
-      - name: api_base
-        dtype: string
-      - name: extra_body
-        struct:
-        - name: chat_template_kwargs
-          struct:
-          - name: enable_thinking
-            dtype: bool
-      - name: max_tokens
-        dtype: int64
-      - name: messages
-        list:
-        - name: content
-          dtype: string
-        - name: role
-          dtype: string
-      - name: model
-        dtype: string
-      - name: stream
-        dtype: bool
-      - name: temperature
-        dtype: float64
-      - name: timeout
-        dtype: float64
-      - name: top_p
-        dtype: float64
-  - name: response__metadata
-    list:
-    - name: cost
-      dtype: float64
-    - name: finish_reason
-      dtype: string
-    - name: input_tokens
-      dtype: int64
-    - name: model
-      dtype: string
-    - name: output_tokens
-      dtype: int64
-    - name: raw_response
-      struct:
-      - name: choices
-        list:
-        - name: finish_reason
-          dtype: string
-        - name: index
-          dtype: int64
-        - name: message
-          struct:
-          - name: content
-            dtype: string
-          - name: function_call
-            dtype: 'null'
-          - name: provider_specific_fields
-            struct:
-            - name: reasoning
-              dtype: 'null'
-            - name: refusal
-              dtype: 'null'
-          - name: role
-            dtype: string
-          - name: tool_calls
-            dtype: 'null'
-        - name: provider_specific_fields
-          struct:
-          - name: seed
-            dtype: 'null'
-      - name: created
-        dtype: int64
-      - name: id
-        dtype: string
-      - name: metadata
-        struct:
-        - name: weight_version
-          dtype: string
-      - name: model
-        dtype: string
-      - name: object
-        dtype: string
-      - name: prompt
-        list: 'null'
-      - name: service_tier
-        dtype: 'null'
-      - name: system_fingerprint
-        dtype: 'null'
-      - name: usage
-        struct:
-        - name: completion_tokens
-          dtype: int64
-        - name: completion_tokens_details
-          dtype: 'null'
-        - name: prompt_tokens
-          dtype: int64
-        - name: prompt_tokens_details
-          dtype: 'null'
-        - name: total_tokens
-          dtype: int64
-      - name: warnings
-        list:
-        - name: message
-          dtype: string
-  - name: response__extractions
-    list:
-      list: string
-  - name: response__extractions_metadata
-    list:
-      list:
-      - name: end_index
-        dtype: int64
-      - name: end_phrase
-        dtype: string
-      - name: start_index
-        dtype: int64
-      - name: start_phrase
-        dtype: string
-  - name: eval_correct
-    list:
-      list: bool
-  - name: eval_correct__metadata
-    list:
-      list:
-      - name: answer_block
-        dtype: string
-      - name: error
-        dtype: string
-      - name: final_answer
-        dtype: float64
-      - name: is_correct
-        dtype: bool
-  splits:
-  - name: train
-    num_bytes: 306524
-    num_examples: 10
-  download_size: 139005
-  dataset_size: 306524
-configs:
-- config_name: default
-  data_files:
-  - split: train
-    path: data/train-*
+license: mit
+tags:
+- t1_synthesize_knowledge
+- structured_reasoning
+- countdown
+- comparison
 ---
+
+# t1-structured-reasoning-full-kimi-k2-instruct-0905
+
+Structured reasoning evaluation: instead of injecting synthesized facts, this uses a
+**static system prompt** that teaches the model a heuristic search FORMAT with explicit
+structural markers ([STEP], [PRUNE], [BACKTRACK], [REVIEW OPTIONS], [SOLUTION FOUND]).
+
+Inspired by HandCraftedCountdownSearch — models SFT'd on structured search traces
+significantly outperform free-form CoT. This tests whether prompt-time format instruction
+alone can capture some of that benefit.
+
+## Performance Comparison
+
+![Comparison Chart](figures/comparison_chart.png)
+
+| Metric | Base | Structured | Delta |
+|--------|------|------------|-------|
+| pass@1 | 0.9000 (90.0%) | 0.3000 (30.0%) | -0.6000 (-60.0%) |
+
+
+## Experimental Details
+
+| Parameter | Value |
+|-----------|-------|
+| Inference model | `together_ai/moonshotai/Kimi-K2-Instruct-0905` |
+| Prompt style | `full` |
+| Temperature | 0.7 |
+| Top-p | 0.95 |
+| Max tokens | 65536 |
+| Samples per problem | 1 |
+| Enable thinking | True |
+| Total problems | 10 |
+| Base eval source | `t1-base-eval-together_ai-qwen-qwen3-235b-a22b-thinkin-12arg` |
+
+## System Prompt Used (full style)
+
+```
+You are solving a countdown arithmetic problem. You must find an expression using the given numbers (each exactly once) that equals the target.
+
+REASON USING STRUCTURED HEURISTIC SEARCH. Do not just guess — systematically search for the answer using the format below. This is critical: use the exact markers shown to organize your reasoning.
+
+=== REASONING FORMAT ===
+
+Use these markers to structure every step of your thinking:
+
+[STEP N: COMBINE] — Try combining two numbers with an operation. Show: a OP b = result. State remaining numbers and how far from target.
+
+[PRUNE] — BEFORE trying an operation, check if it's useful. Skip it if:
+  - Division doesn't produce an integer
+  - The result moves AWAY from target (current < target? don't subtract/divide. current > target? don't add/multiply)
+  Say what you're skipping and why.
+
+[BACKTRACK] — Current path is stuck or getting worse. State why (too high? too low? no numbers left?) and go back to try a different combination.
+
+[REVIEW OPTIONS] — When stuck, list ALL remaining possible combinations, score each by |result - target|, and pick the best one to try next.
+
+[SOLUTION FOUND] — You hit the target! Verify: recompute the expression, confirm each number is used exactly once, then give the answer.
+
+=== KEY HEURISTIC ===
+
+DIRECTIONAL SEARCH: At every step, check if your current value is above or below the target.
+  - Below target → only try + and * (go UP)
+  - Above target → only try - and / (go DOWN)
+This cuts your search space in half and prevents wasted exploration.
+
+=== WORKED EXAMPLE ===
+
+Target: 23, Numbers: [4, 7, 3]
+
+[STEP 1: COMBINE]
+Let me scan first moves. Best candidates by closeness to 23:
+  4 * 7 = 28 (distance 5) ← closest
+  4 + 7 = 11 (distance 12)
+  7 * 3 = 21 (distance 2) ← very close!
+  7 + 3 = 10 (distance 13)
+Try: 7 * 3 = 21. Remaining: [4]. Current: 21, need 23.
+
+[STEP 2: COMBINE]
+Current 21 < target 23, so only try + and *.
+21 + 4 = 25 (distance 2)
+21 * 4 = 84 (distance 61)
+[PRUNE] 21 * 4 = 84 — way t...
+```
+
+## Dataset Info
+
+- **Rows**: Structured reasoning evaluation results (one row per problem)
+- **Columns**: question, metadata, task_source, formatted_prompt, response, eval_correct, and associated metadata
+
+## Usage
+
+```python
+from datasets import load_dataset
+
+dataset = load_dataset("reasoning-degeneration-dev/t1-structured-reasoning-full-kimi-k2-instruct-0905", split="train")
+print(f"Loaded {len(dataset)} rows")
+```
+
+---
+
+*This dataset is tracked in [reasoning-degeneration-dev/PROJECT-MANIFEST](https://huggingface.co/datasets/reasoning-degeneration-dev/PROJECT-MANIFEST)*