Add paper link and task category

#1
by nielsr HF Staff - opened
Files changed (1) hide show
  1. README.md +15 -14
README.md CHANGED
@@ -1,28 +1,29 @@
1
  ---
2
  license: apache-2.0
 
 
3
  ---
4
 
5
- Datasets and query sets for the three benchmark settings from the "Generating Expressive and Customizable Evals for Timeseries Data Analysis Agents with AgentFuel" paper.
 
 
6
 
7
  All datasets and query sets were generated using AgentFuel's data generation and question-answer generation modules.
8
 
9
- E-commerce
10
  Product analytics for an e-commerce website. Browsing sessions are generated using a state machine covering browsing flows, cart abandonment, and purchase flows.
11
 
12
- Datasets: ecommerce_users_data.csv, ecommerce_sessions_data.csv
13
-
14
- Queries: 12 stateless (ecommerce_basic.csv) + 12 stateful (ecommerce_stateful.csv)
15
 
16
- IoT
17
  IoT device monitoring with three sensor exemplars: temperature, pressure, and humidity, each with its own operations state machine and device health metrics.
18
 
19
- Dataset: iot_device_data.csv
20
-
21
- Queries: 12 stateless (iot_basic.csv) + 12 stateful (iot_stateful.csv)
22
-
23
- Telecom
24
- Telecommunications network telemetry across three related entities: cell sites, transport links, and core nodes. The _with_inc_ dataset variants include an injected cascading incident: a transport link degrades (elevated packet loss, latency, jitter), cascading to connected cell sites (higher RRC failures, lower availability), with a modest effect on core nodes (reduced attached UEs, increased CPU load).
25
 
26
- Datasets: cell_site_data.csv, transport_link_data.csv, core_node_data.csv (and _with_inc_ variants)
 
27
 
28
- Queries: 12 stateless (telecom_basic.csv) + 12 incident-specific (telecom_incident.csv)
 
 
1
  ---
2
  license: apache-2.0
3
+ task_categories:
4
+ - text-generation
5
  ---
6
 
7
+ # AgentFuel Benchmarks
8
+
9
+ Datasets and query sets for the three benchmark settings from the paper [Generating Expressive and Customizable Evals for Timeseries Data Analysis Agents with AgentFuel](https://huggingface.co/papers/2603.12483).
10
 
11
  All datasets and query sets were generated using AgentFuel's data generation and question-answer generation modules.
12
 
13
+ ### E-commerce
14
  Product analytics for an e-commerce website. Browsing sessions are generated using a state machine covering browsing flows, cart abandonment, and purchase flows.
15
 
16
+ - **Datasets**: `ecommerce_users_data.csv`, `ecommerce_sessions_data.csv`
17
+ - **Queries**: 12 stateless (`ecommerce_basic.csv`) + 12 stateful (`ecommerce_stateful.csv`)
 
18
 
19
+ ### IoT
20
  IoT device monitoring with three sensor exemplars: temperature, pressure, and humidity, each with its own operations state machine and device health metrics.
21
 
22
+ - **Dataset**: `iot_device_data.csv`
23
+ - **Queries**: 12 stateless (`iot_basic.csv`) + 12 stateful (`iot_stateful.csv`)
 
 
 
 
24
 
25
+ ### Telecom
26
+ Telecommunications network telemetry across three related entities: cell sites, transport links, and core nodes. The `_with_inc_` dataset variants include an injected cascading incident: a transport link degrades (elevated packet loss, latency, jitter), cascading to connected cell sites (higher RRC failures, lower availability), with a modest effect on core nodes (reduced attached UEs, increased CPU load).
27
 
28
+ - **Datasets**: `cell_site_data.csv`, `transport_link_data.csv`, `core_node_data.csv` (and `_with_inc_` variants)
29
+ - **Queries**: 12 stateless (`telecom_basic.csv`) + 12 incident-specific (`telecom_incident.csv`)