agentabstain

KO | EN

gitlite — search

#gpt #devops #vue #web #privacy #ios #hacktoberfest #list #cpp #csharp

(★ 9)

AgentAbstain: Systematic Evaluation Framework of Agentic Abstention

File Explorer

ZIP Download

- - __init__.py
- - __init__.py
  - base.py
  - multi.py
  - registry.py
- - __init__.py
- __init__.py
- - __init__.py
  - agent.py
  - mcp_bridge.py
- - __init__.py
  - agent.py
  - toolset.py
- - __init__.py
  - agent.py
  - mcp_server.py
- - __init__.py
  - agent.py
  - README.md
- __init__.py
- figure1.png
- logo.png
- ranking_bar.png
- - default.yaml
- - __init__.py
  - base.py
  - commit_check.py
  - response_llm_judge.py
- - eval.sh
- - __init__.py
  - analysis.py
  - figure_act_abstain_dumbbell.py
  - figure_act_vs_abstain_centroid.py
  - figure_category_difficulty.py
  - figure_failure_modes_2x2.py
  - figure_failure_modes_bar.py
  - figure_leaderboard_bar.py
  - figure_ranking_bar.py
  - figure_scaling_lineplot.py
- __init__.py
- config.py
- discovery.py
- loader.py
- runner.py
- - claudesdk_claude-haiku-4-5.yaml
  - claudesdk_claude-opus-4-7.yaml
  - claudesdk_claude-sonnet-4-6.yaml
  - googleadk_gemini-3-flash-preview.yaml
  - googleadk_gemini-3.1-pro.yaml
  - openaisdk_gpt-4o.yaml
  - openaisdk_gpt-5.1.yaml
  - openaisdk_gpt-5.2.yaml
  - openaisdk_gpt-5.4.yaml
  - openaisdk_gpt-5.5.yaml
  - openaisdk_gpt-5.yaml
  - openclaw_deepseek-v3-2.yaml
  - openclaw_deepseek-v4-pro.yaml
  - openclaw_glm-5.yaml
  - openclaw_gpt-oss-120b.yaml
  - openclaw_kimi-k2-5.yaml
  - openclaw_minimax-m2-5.yaml
  - tasks.yaml
  - tasks_high_stakes_action.yaml
  - tasks_insufficient_tool_capability.yaml
  - tasks_preview.yaml
  - tasks_smoke.yaml
- - __init__.py
  - claudesdk.py
  - common.py
  - config.py
  - googleadk.py
  - openaisdk.py
  - openclaw.py
  - task_mcp_server.py
- - __init__.py
  - resume_session.py
  - run.sh
  - run_inference.py
  - run_new_categories.sh
- - __init__.py
  - BaseAgent.py
  - trajectory.py
- __init__.py
.env.template
.gitignore
LICENSE
README.md
requirements.txt

# Use via CDN

jsDelivr

jsDelivr serves any public GitHub repository as a CDN with zero setup. Pick a version and a file to get a ready-to-paste link and snippet.

Version

File

Link

Example

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

<h1>AgentAbstain: Do LLM Agents Know When Not to Act?</h1>

<p><i>The first systematic evaluation framework for agentic abstention:<br>the calibrated ability of tool-using LLM agents to recognize when <b>not</b> to act.</i></p>
</div>

---

## Overview

Most agent benchmarks reward getting the task done. AgentAbstain tests the opposite skill: knowing when to stop. When a request is vague, self-contradictory, or impossible with the tools on hand, an agent that plows ahead can do real and irreversible damage.

The framework has two components:

- **AgentAbstain**, the benchmark: **263 paired tasks** in **42 executable MCP sandbox environments**, built on an agent-native taxonomy of **8 abstention scenarios** spanning pre-execution and runtime triggers. Every should-act task ships with a should-abstain variant that differs by a single controlled perturbation, so no always-act or always-refuse policy can exceed 50% paired accuracy.
- **AbstainGen**, the pipeline: synthesizes the environments and task pairs end-to-end, validated by deterministic DAG replay and cross-family LLM critics. Three independent annotators rate 94 to 98% of sampled tasks as well-designed.

Evaluation crosses a deterministic **commit check** on the tool-call trace with an **LLM judge** on the terminal response, which isolates failure modes that neither signal catches alone, such as *post-hoc abstention*: the agent executes the irreversible action first and claims restraint afterwards.

<div align="center">
  <img src="assets/figure1.png" width="620" alt="One should-abstain task, four qualitatively different agent behaviors">
  <p><i>One should-abstain task, four qualitatively different behaviors. Only Successful Abstention, where the agent holds the critical action and surfaces the conflict, counts as correct.</i></p>
</div>

## Leaderboard

Across 17 frontier LLMs in 4 agent harnesses, the best agent reaches only **59.5% paired accuracy**, and abstention capability is largely independent of task-solving capability.

| # | Model | Harness | Act | Abstain | Paired | CAR |
|--:|:------|:--------|----:|--------:|-------:|----:|
| 1 | Gemini 3.1 Pro | Google ADK | 90.5 | 65.4 | **59.5** | 65.7 |
| 2 | Claude Opus 4.7 | Claude SDK | 76.5 | **79.0** | 59.4 | **77.6** |
| 3 | Claude Sonnet 4.6 | Claude SDK | 83.1 | 66.4 | 53.4 | 65.4 |
| 4 | GPT-5.5 | OpenAI SDK | 87.4 | 61.1 | 52.5 | 59.8 |
| 5 | Claude Haiku 4.5 | Claude SDK | 80.5 | 65.6 | 49.7 | 61.8 |
| 6 | GPT-5 | OpenAI SDK | 74.6 | 69.8 | 49.6 | 66.5 |
| 7 | GPT-5.4 | OpenAI SDK | 76.9 | 67.8 | 48.7 | 64.0 |
| 8 | GLM-5 | OpenClaw | 82.5 | 61.8 | 47.8 | 59.1 |
| 9 | GPT-OSS 120B | OpenClaw | 78.3 | 59.5 | 46.2 | 58.2 |
| 10 | GPT-5.2 | OpenAI SDK | 74.7 | 63.1 | 42.9 | 59.2 |
| 11 | MiniMax M2.5 | OpenClaw | 83.8 | 50.1 | 41.9 | 49.6 |
| 12 | DeepSeek V3.2 | OpenClaw | 82.4 | 52.1 | 41.4 | 50.2 |
| 13 | GPT-5.1 | OpenAI SDK | 75.0 | 60.7 | 40.6 | 53.2 |
| 14 | Gemini 3 Flash | Google ADK | **91.7** | 43.6 | 39.7 | 43.4 |
| 15 | DeepSeek V4 Pro | OpenClaw | 87.0 | 42.8 | 36.9 | 42.3 |
| 16 | Kimi K2.5 | OpenClaw | 63.9 | 52.0 | 33.4 | 53.2 |
| 17 | GPT-4o | OpenAI SDK | 82.1 | 44.2 | 33.0 | 40.9 |

**Paired** (primary): the share of pairs where the model gets both the should-act and should-abstain variants right. **Act** and **Abstain**: per-side pass rates. **CAR** (Conditioned Abstention Rate): abstain accuracy restricted to pairs whose act side the model already solved, isolating restraint from raw capability. All numbers are macro-averaged over the 8 scenarios.

## Repository Layout

```
agent/                        harness adapters (Claude SDK, OpenAI SDK, Google ADK, OpenClaw)
abstention_factory/           vendored runtime core: environment contract, registry, shared utils
                              (the AbstainGen generation pipeline itself is not released)

src/                          inference runtime
├── runtime/                  agent harness integrations (Claude SDK, OpenAI SDK, Google ADK, OpenClaw)
├── configs/                  17 model configs + task sets (tasks.yaml is the full benchmark)
├── scripts/                  run.sh, run_inference.py, resume_session.py
└── types/                    shared task and rollout types

eval/                         evaluation harness
├── evaluators/               commit check (deterministic) + LLM response judge
├── configs/default.yaml      judge configuration
├── runner.py                 per-model evaluation entry point
├── statistics/               analysis and figure scripts behind every figure in the paper
└── scripts/eval.sh           batch evaluation across models
```

## Quick Start

### 1. Set up an environment

```bash
git clone https://github.com/AntiQuality/agentabstain && cd agentabstain

conda create -n agentabstain python=3.11 -y
conda activate agentabstain
pip install -r requirements.txt

# or with uv
uv venv --python 3.11 && source .venv/bin/activate
uv pip install -r requirements.txt
```

### 2. Fetch the benchmark

```bash
# 263 task pairs + 42 executable environments, from Hugging Face
python -c "from huggingface_hub import snapshot_download; \
           snapshot_download('antiquality/agentabstain', repo_type='dataset', local_dir='data')"
```

The runtime reads tasks and environments from `./data` by default; set `AGENTABSTAIN_DATA` to use another location.

### 3. Configure credentials

```bash
cp .env.template .env    # then fill in the keys you need
```

Inference and evaluation both load `.env` automatically; values already exported in your shell win. You only need the keys for what you run: `OPENAI_API_KEY` powers the OpenAI SDK harness and the response judge (any OpenAI-compatible gateway works via `OPENAI_BASE_URL`), `GOOGLE_API_KEY` the Google ADK harness, and `OPENROUTER_API_KEY` the OpenClaw harness. The Claude SDK harness takes either `ANTHROPIC_API_KEY` with native model IDs or the Bedrock route pinned in the shipped configs; the template documents both.

### 4. Smoke test

Verify the full loop on a single task pair (two runs, a few cents of API usage). Works with any model config:

```bash
python -m src.scripts.run_inference \
    --runtime-config src/configs/openclaw_deepseek-v4-pro.yaml --smoke
python -m eval.runner --provider openclaw --model openrouter/deepseek/deepseek-v4-pro
```

OpenClaw-harness models additionally need the openclaw CLI pinned to the version used for the paper's evaluation campaign (`npm install -g openclaw@2026.4.29`; Node 22.14 to 23.x, since this older build predates Node 24 native-module ABIs; see `agent/openclaw/README.md`). Newer openclaw releases (2026.7+) changed the agent config schema and also alter how the per-task system prompt reaches the model, so they are not drop-in compatible with this harness.

### 5. Run inference

Rollouts are written to `results/{provider}/{model}/`:

```bash
# all models of one provider family
bash src/scripts/run.sh claudesdk        # openaisdk | googleadk | claudesdk | all

# or one model against one task set
python -m src.scripts.run_inference \
    --runtime-config src/configs/claudesdk_claude-opus-4-7.yaml \
    --task-config    src/configs/tasks.yaml \
    --workers 4
```

### 6. Run evaluation

The commit check and the LLM judge score saved rollouts; the judge is configured in `eval/configs/default.yaml`. If your gateway namespaces model IDs (OpenRouter, for example, wants `openai/gpt-5.4` rather than the bare OpenAI ID), change `judge_models[].model` there to match. `--model` is the model string from the runtime config, i.e. the directory name under `results/{provider}/`:

```bash
python -m eval.runner --provider claudesdk --model us.anthropic.claude-opus-4-7

# or batch across models
bash eval/scripts/eval.sh
```

Judge verdicts are cached in each run's `eval.json` and reused on re-runs, including error records from a misconfigured judge; pass `--override-judge` to re-judge after fixing the configuration.

### 7. Regenerate figures

Every figure in the paper is produced by a script under `eval/statistics/` (for example `figure_ranking_bar.py`, `figure_category_difficulty.py`); each reads the evaluation outputs from your own runs.

## Dataset

The 263 task pairs and the 42 sandbox environments are hosted on [Hugging Face](https://huggingface.co/datasets/antiquality/agentabstain); task sets under `src/configs/` reference dataset task IDs. The AbstainGen generation pipeline is fully documented in the paper and intentionally not open-sourced, as a public generator would let benchmark-matched training data be synthesized at scale. Fresh evaluation rounds can be generated privately on demand, which keeps the benchmark resistant to training-data contamination.

## License

Code is released under the [MIT License](LICENSE). The [dataset](https://huggingface.co/datasets/antiquality/agentabstain) is released under CC BY 4.0.

## Citation

```bibtex
@misc{liu2026agentabstain,
  title  = {AgentAbstain: Do LLM Agents Know When Not to Act?},
  author = {Liu, Xun and Zhang, Yi Evie and Kasprova, Vira and Rabbani, Parisa and Zahraei, Pardis Sadat and Zhang, Tianyu and Ebrahimpour-Boroojeny, Ali and Chandrasekaran, Varun},
  year   = {2026},
  eprint = {2607.10059},
  archivePrefix = {arXiv},
  primaryClass  = {cs.AI}
}
```

agentabstain

Owner Repos

File Explorer

# Use via CDN

Was this content helpful?

Recent Feedback

agentabstain

Owner Repos

File Explorer

# Use via CDN

Was this content helpful?

Recent Feedback

Download README