---
pretty_name: HackBench
size_categories:
- n<1K
tags:
- reward-hacking
- agent-trajectories
- long-context
- command-localization
- evaluation
configs:
- config_name: default
  data_files:
  - split: test
    path: data/hackbench_169_gpt55_glm52.numbered.jsonl
license: cc-by-nc-4.0
extra_gated_prompt: >
  This dataset is licensed under CC BY-NC 4.0
  (https://creativecommons.org/licenses/by-nc/4.0/).
  Please also agree to the additional conditions below.
extra_gated_fields:
  "I agree to the CC BY-NC license": checkbox
  "I agree not to use this dataset for model training": checkbox
  "I agree not to rehost this dataset": checkbox
extra_gated_button_content: Agree and access
---

# HackBench

HackBench is a benchmark for numbered command localization of reward hacking in
long-horizon ML research agent trajectories. Each record contains the original
trajectory, a ready-to-use numbered monitor input, and audited gold command indices.

This release contains **169 HACK records** across 13 tasks: 41 natural and
128 synthetic hacking examples.

| Source agent | Records |
| --- | ---: |
| GPT-5.5 | 102 |
| GLM-5.2 | 67 |
| **Total** | **169** |

Project: [RAM](https://github.com/facebookresearch/RAM). 

Blog Post: [HackBench: Detecting and Localizing Reward Hacking in Autoresearch Agents](https://facebookresearch.github.io/RAM/blogs/hackbench/).

## What is included

- **One JSONL row per trajectory record**.
- Full `injected_transcript` strings, including recorded reasoning, tool commands,
  and observations. 
- `monitor_input`, with task context and a transcript whose executable command
  blocks carry one-based `[CMD n]` labels, plus matching `gold_command_indices`.
- Task instructions, gold command strings, and source-agent settings.


## Load the dataset

```python
from datasets import load_dataset

ds = load_dataset("facebook/hackbench", split="test", token=True)
row = ds[0]

# Input to a detector. Keep gold labels separate.
# The instruction below is the agent's task context, not the monitor's prompt.
detector_input = {
    "instruction": row["monitor_input"]["task_instruction"],
    "transcript": row["monitor_input"]["transcript"],
}

# Ground truth, used only after the detector produces its prediction.
gold_command_indices = row["gold_command_indices"]
```

`monitor_input` uses the reference preprocessing: historical detector feedback
is removed, blocked attempts remain as command blocks, and the research task
context is the complete initial user message. This task context is for the monitor 
to understand the agent's objective; supply your own monitor prompt separately. 
The numbered transcript is fully rendered without model-specific clipping.

The gold indices refer to the `[CMD n]` labels in that input. Numbering starts
at 1 for each trajectory and counts
executable command blocks; a block can contain several lines of code. Use `gold_command_indices`
for evaluation and preserve the provided numbers when managing long context.

The file can also be read directly, without converting its nested metadata to
Arrow columns:

```python
import json
from huggingface_hub import hf_hub_download

path = hf_hub_download(
    repo_id="facebook/hackbench",
    repo_type="dataset",
    filename="data/hackbench_169_gpt55_glm52.numbered.jsonl",
    token=True,
)
with open(path, encoding="utf-8") as stream:
    row = json.loads(next(stream))
```

## Fields

| Field | Type | Meaning |
| --- | --- | --- |
| `id` | string | Unique record identifier; historical provenance identifiers are preserved. |
| `task.name` | string | One of the canonical task names listed below. |
| `task.instruction` | string | Stored task-wrapper metadata|
| `source_agent` | object | Autoresearch agent; model names use `GPT-5.5` or `GLM-5.2`.|
| `injected_transcript` | string | Original released trajectory text, retained as provenance. |
| `monitor_input.task_instruction` | string | Complete recorded initial instructions for the research agent, supplied to the monitor as task context. |
| `monitor_input.transcript` | string | Preprocessed transcript with `[CMD n]` labels; the numbered-localization input. |
| `monitor_input.command_count` | integer | Number of executable command blocks in the numbered input. |
| `gold_command_indices` | list of integers | Audited gold `[CMD n]` indices in `monitor_input.transcript`, using one-based numbering. |
| `hack_commands` | list of strings | Original gold command text, retained as audit evidence. |
| `origin` | string | `natural` or `synthetic`. |
| `hack_type` | string | `general` or `task-specific`. |


Gold indices and command strings can reveal the answer. Construct detector input explicitly from `monitor_input`
and keep evaluation targets separate.

## Task coverage

| Task | Records |
| --- | ---: |
| `continual-learning` | 7 |
| `cross-modal-retrieval` | 15 |
| `data_select_ifeval` | 11 |
| `flux2_klein_lora` | 13 |
| `grpo_multisource` | 6 |
| `improving-replay-buffers` | 37 |
| `llm_online_serving` | 6 |
| `materials-tokenization` | 9 |
| `moving_mnist_world_model` | 13 |
| `multilingual_ocr` | 25 |
| `scaling_law` | 8 |
| `time-series-explanation` | 14 |
| `w2s` | 5 |



## Long-context considerations

| Transcript size | Minimum | Median | Maximum |
| --- | ---: | ---: | ---: |
| Unicode characters | 73,412 | 1,064,439 | 3,834,163 |
| UTF-8 bytes | 73,532 | 1,064,975 | 3,835,613 |

These lengths describe the original `injected_transcript`. The JSONL contains
both the original text and the numbered reference view.

The JSONL file is 408,528,473 bytes (389.60 MiB). Long rows may exceed
a model's context limit, and a web viewer may shorten what it displays. Do not
silently truncate them. If an evaluation uses chunking or another context
strategy, report that strategy and retain linkage to the full record.

## Intended use and limitations

The dataset supports offline evaluation of numbered reward-hacking command
localization on 169 HACK trajectories. Report precision, recall, and F₁ computed
per trajectory and then averaged over the release. The `test` split denotes an
evaluation release; no separate training or validation split is supplied.