exp_name stringclasses 3
values | n_tokens_action int64 27 587 | system_prompt stringclasses 1
value | user_prompt stringlengths 5.59k 79.8k | output stringlengths 112 2.26k | n_tokens int64 1.55k 20.8k | task_name stringclasses 24
values | agent_args_dict dict | reward float64 0.74 0.98 | step_idx int64 0 14 | total_steps int64 1 15 | goal stringclasses 132
values | log_probs null | tokens null | repetition_penalty int64 0 0 | seed int64 4 996 | og_reward int64 1 1 | trace_idx int64 3 5.93k | split stringclasses 1
value | agent_args dict | text stringlengths 6.54k 80.8k | spans listlengths 1 1 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
KIMI | 80 | You are an agent trying to solve a web task based on the content of the page and
user instructions. You can interact with the page and explore, and send messages to the user. Each time you
submit an action it will be sent to the browser and you will receive a new page. | # Instructions
Review the current state of the page and all other information to find the best
possible next action to accomplish your goal. Your answer will be interpreted
and executed by a program, make sure to follow the formatting instructions.
## Goal:
What is the value of "Bess Marso" in the chart (in count)?
... |
<think>
I need to find the value of "Bess Marso" in the chart. The current page shows a loading report, so I need to wait for it to load and then look for the chart data. Since the page is still loading, I'll wait a moment for the content to appear.
</think>
<action>
noop(2000)
</action>
| 1,990 | workarena.servicenow.single-chart-value-retrieval | {
"agent_name": "GenericAgent-moonshotai_kimi-k2",
"chat_model_args": {
"log_probs": false,
"max_input_tokens": 100000,
"max_new_tokens": 128000,
"max_total_tokens": 128000,
"model_name": "moonshotai/kimi-k2",
"temperature": 0.1,
"vision_support": false
},
"flags": {
"action": {
... | 0.9604 | 0 | 2 | What is the value of "Bess Marso" in the chart (in count)? | null | null | 0 | 661 | 1 | 3 | train | {
"agent_name": "GenericAgent-moonshotai_kimi-k2",
"chat_model_args": {
"log_probs": false,
"max_input_tokens": 100000,
"max_new_tokens": 128000,
"max_total_tokens": 128000,
"model_name": "moonshotai/kimi-k2",
"temperature": 0.1,
"vision_support": false
},
"flags": {
"action": {
... | <s><|system|>
You are a thoughtful and systematic AI assistant built by ServiceNow Language Models (SLAM) lab. Before providing an answer, analyze the problem carefully and present your reasoning step by step. After explaining your thought process, provide the final solution in the following format: [BEGIN FINAL RESPON... | [
[
0,
7637
]
] |
KIMI | 169 | You are an agent trying to solve a web task based on the content of the page and
user instructions. You can interact with the page and explore, and send messages to the user. Each time you
submit an action it will be sent to the browser and you will receive a new page. | # Instructions
Review the current state of the page and all other information to find the best
possible next action to accomplish your goal. Your answer will be interpreted
and executed by a program, make sure to follow the formatting instructions.
## Goal:
What is the value of "Bess Marso" in the chart (in count)?
... |
<think>
Looking at the current page state, I can see that the report has loaded and there's a chart displayed. The AXTree shows that there's a Highcharts interactive chart with 9 bars. Most importantly, I can see in the SVGRoot section that there's an image element showing "2. Bess Marso, 21, 100.0%". This indicates t... | 2,718 | workarena.servicenow.single-chart-value-retrieval | {
"agent_name": "GenericAgent-moonshotai_kimi-k2",
"chat_model_args": {
"log_probs": false,
"max_input_tokens": 100000,
"max_new_tokens": 128000,
"max_total_tokens": 128000,
"model_name": "moonshotai/kimi-k2",
"temperature": 0.1,
"vision_support": false
},
"flags": {
"action": {
... | 0.9604 | 1 | 2 | What is the value of "Bess Marso" in the chart (in count)? | null | null | 0 | 661 | 1 | 3 | train | {
"agent_name": "GenericAgent-moonshotai_kimi-k2",
"chat_model_args": {
"log_probs": false,
"max_input_tokens": 100000,
"max_new_tokens": 128000,
"max_total_tokens": 128000,
"model_name": "moonshotai/kimi-k2",
"temperature": 0.1,
"vision_support": false
},
"flags": {
"action": {
... | <s><|system|>
You are a thoughtful and systematic AI assistant built by ServiceNow Language Models (SLAM) lab. Before providing an answer, analyze the problem carefully and present your reasoning step by step. After explaining your thought process, provide the final solution in the following format: [BEGIN FINAL RESPON... | [
[
0,
10028
]
] |
KIMI | 241 | You are an agent trying to solve a web task based on the content of the page and
user instructions. You can interact with the page and explore, and send messages to the user. Each time you
submit an action it will be sent to the browser and you will receive a new page. | # Instructions
Review the current state of the page and all other information to find the best
possible next action to accomplish your goal. Your answer will be interpreted
and executed by a program, make sure to follow the formatting instructions.
## Goal:
Create a new change request with a value of "zootomical coll... |
<think>
I need to create a new change request with specific values for several fields. Looking at the current page, I can see the form is already open with the change request number CHG0031164. I need to fill in the required fields:
1. Category: Should be "Network" (currently set to "Other")
2. Configuration item: Sh... | 4,850 | workarena.servicenow.create-change-request | {
"agent_name": "GenericAgent-moonshotai_kimi-k2-0905",
"chat_model_args": {
"log_probs": false,
"max_input_tokens": 100000,
"max_new_tokens": 128000,
"max_total_tokens": 128000,
"model_name": "moonshotai/kimi-k2-0905",
"temperature": 0.1,
"vision_support": false
},
"flags": {
"a... | 0.833748 | 0 | 9 | Create a new change request with a value of "zootomical collumelliaceous saleratus enolizable Rudolph" for field "Close notes", a value of "High" for field "Risk", a value of "lnux101" for field "Configuration item", a value of "Network" for field "Category", a value of "We need to add additional redundant name servers... | null | null | 0 | 475 | 1 | 4 | train | {
"agent_name": "GenericAgent-moonshotai_kimi-k2-0905",
"chat_model_args": {
"log_probs": false,
"max_input_tokens": 100000,
"max_new_tokens": 128000,
"max_total_tokens": 128000,
"model_name": "moonshotai/kimi-k2-0905",
"temperature": 0.1,
"vision_support": false
},
"flags": {
"a... | <s><|system|>
You are a thoughtful and systematic AI assistant built by ServiceNow Language Models (SLAM) lab. Before providing an answer, analyze the problem carefully and present your reasoning step by step. After explaining your thought process, provide the final solution in the following format: [BEGIN FINAL RESPON... | [
[
0,
17098
]
] |
KIMI | 91 | You are an agent trying to solve a web task based on the content of the page and
user instructions. You can interact with the page and explore, and send messages to the user. Each time you
submit an action it will be sent to the browser and you will receive a new page. | # Instructions
Review the current state of the page and all other information to find the best
possible next action to accomplish your goal. Your answer will be interpreted
and executed by a program, make sure to follow the formatting instructions.
## Goal:
Create a new change request with a value of "zootomical coll... |
<think>
I can see that the Category field is already set to "Network" which is correct. Now I need to set the Risk field to "High". Looking at the AXTree, I can see the Risk combobox with bid 'a342' currently set to "-- None --". I need to change this to "High".
</think>
<action>
select_option('a342', 'High')
</actio... | 5,139 | workarena.servicenow.create-change-request | {
"agent_name": "GenericAgent-moonshotai_kimi-k2-0905",
"chat_model_args": {
"log_probs": false,
"max_input_tokens": 100000,
"max_new_tokens": 128000,
"max_total_tokens": 128000,
"model_name": "moonshotai/kimi-k2-0905",
"temperature": 0.1,
"vision_support": false
},
"flags": {
"a... | 0.833748 | 1 | 9 | Create a new change request with a value of "zootomical collumelliaceous saleratus enolizable Rudolph" for field "Close notes", a value of "High" for field "Risk", a value of "lnux101" for field "Configuration item", a value of "Network" for field "Category", a value of "We need to add additional redundant name servers... | null | null | 0 | 475 | 1 | 4 | train | {
"agent_name": "GenericAgent-moonshotai_kimi-k2-0905",
"chat_model_args": {
"log_probs": false,
"max_input_tokens": 100000,
"max_new_tokens": 128000,
"max_total_tokens": 128000,
"model_name": "moonshotai/kimi-k2-0905",
"temperature": 0.1,
"vision_support": false
},
"flags": {
"a... | <s><|system|>
You are a thoughtful and systematic AI assistant built by ServiceNow Language Models (SLAM) lab. Before providing an answer, analyze the problem carefully and present your reasoning step by step. After explaining your thought process, provide the final solution in the following format: [BEGIN FINAL RESPON... | [
[
0,
18195
]
] |
KIMI | 99 | You are an agent trying to solve a web task based on the content of the page and
user instructions. You can interact with the page and explore, and send messages to the user. Each time you
submit an action it will be sent to the browser and you will receive a new page. | # Instructions
Review the current state of the page and all other information to find the best
possible next action to accomplish your goal. Your answer will be interpreted
and executed by a program, make sure to follow the formatting instructions.
## Goal:
Create a new change request with a value of "zootomical coll... |
<think>
I can see that the Risk field has been successfully set to "High". Now I need to fill in the Configuration item field with "lnux101". Looking at the AXTree, I can see the Configuration item searchbox with bid 'a300' is currently empty. I need to fill this with "lnux101".
</think>
<action>
fill('a300', 'lnux10... | 5,247 | workarena.servicenow.create-change-request | {
"agent_name": "GenericAgent-moonshotai_kimi-k2-0905",
"chat_model_args": {
"log_probs": false,
"max_input_tokens": 100000,
"max_new_tokens": 128000,
"max_total_tokens": 128000,
"model_name": "moonshotai/kimi-k2-0905",
"temperature": 0.1,
"vision_support": false
},
"flags": {
"a... | 0.833748 | 2 | 9 | Create a new change request with a value of "zootomical collumelliaceous saleratus enolizable Rudolph" for field "Close notes", a value of "High" for field "Risk", a value of "lnux101" for field "Configuration item", a value of "Network" for field "Category", a value of "We need to add additional redundant name servers... | null | null | 0 | 475 | 1 | 4 | train | {
"agent_name": "GenericAgent-moonshotai_kimi-k2-0905",
"chat_model_args": {
"log_probs": false,
"max_input_tokens": 100000,
"max_new_tokens": 128000,
"max_total_tokens": 128000,
"model_name": "moonshotai/kimi-k2-0905",
"temperature": 0.1,
"vision_support": false
},
"flags": {
"a... | <s><|system|>
You are a thoughtful and systematic AI assistant built by ServiceNow Language Models (SLAM) lab. Before providing an answer, analyze the problem carefully and present your reasoning step by step. After explaining your thought process, provide the final solution in the following format: [BEGIN FINAL RESPON... | [
[
0,
18606
]
] |
KIMI | 113 | "You are an agent trying to solve a web task based on the content of the page and\nuser instructions(...TRUNCATED) | "# Instructions\nReview the current state of the page and all other information to find the best\npo(...TRUNCATED) | "\n<think>\nI can see that the Configuration item field has been filled with \"lnux101\" and there's(...TRUNCATED) | 5,598 | workarena.servicenow.create-change-request | {"agent_name":"GenericAgent-moonshotai_kimi-k2-0905","chat_model_args":{"log_probs":false,"max_input(...TRUNCATED) | 0.833748 | 3 | 9 | "Create a new change request with a value of \"zootomical collumelliaceous saleratus enolizable Rudo(...TRUNCATED) | null | null | 0 | 475 | 1 | 4 | train | {"agent_name":"GenericAgent-moonshotai_kimi-k2-0905","chat_model_args":{"log_probs":false,"max_input(...TRUNCATED) | "<s><|system|>\nYou are a thoughtful and systematic AI assistant built by ServiceNow Language Models(...TRUNCATED) | [
[
0,
19733
]
] |
KIMI | 143 | "You are an agent trying to solve a web task based on the content of the page and\nuser instructions(...TRUNCATED) | "# Instructions\nReview the current state of the page and all other information to find the best\npo(...TRUNCATED) | "\n<think>\nI can see that the Configuration item field has been successfully set to \"lnux101\". No(...TRUNCATED) | 5,629 | workarena.servicenow.create-change-request | {"agent_name":"GenericAgent-moonshotai_kimi-k2-0905","chat_model_args":{"log_probs":false,"max_input(...TRUNCATED) | 0.833748 | 4 | 9 | "Create a new change request with a value of \"zootomical collumelliaceous saleratus enolizable Rudo(...TRUNCATED) | null | null | 0 | 475 | 1 | 4 | train | {"agent_name":"GenericAgent-moonshotai_kimi-k2-0905","chat_model_args":{"log_probs":false,"max_input(...TRUNCATED) | "<s><|system|>\nYou are a thoughtful and systematic AI assistant built by ServiceNow Language Models(...TRUNCATED) | [
[
0,
19917
]
] |
KIMI | 120 | "You are an agent trying to solve a web task based on the content of the page and\nuser instructions(...TRUNCATED) | "# Instructions\nReview the current state of the page and all other information to find the best\npo(...TRUNCATED) | "\n<think>\nI can see that the Description field has been successfully filled with the required text(...TRUNCATED) | 5,837 | workarena.servicenow.create-change-request | {"agent_name":"GenericAgent-moonshotai_kimi-k2-0905","chat_model_args":{"log_probs":false,"max_input(...TRUNCATED) | 0.833748 | 5 | 9 | "Create a new change request with a value of \"zootomical collumelliaceous saleratus enolizable Rudo(...TRUNCATED) | null | null | 0 | 475 | 1 | 4 | train | {"agent_name":"GenericAgent-moonshotai_kimi-k2-0905","chat_model_args":{"log_probs":false,"max_input(...TRUNCATED) | "<s><|system|>\nYou are a thoughtful and systematic AI assistant built by ServiceNow Language Models(...TRUNCATED) | [
[
0,
20652
]
] |
KIMI | 132 | "You are an agent trying to solve a web task based on the content of the page and\nuser instructions(...TRUNCATED) | "# Instructions\nReview the current state of the page and all other information to find the best\npo(...TRUNCATED) | "\n<think>\nI can see that the \"Risk and impact analysis\" field has been successfully filled with (...TRUNCATED) | 5,862 | workarena.servicenow.create-change-request | {"agent_name":"GenericAgent-moonshotai_kimi-k2-0905","chat_model_args":{"log_probs":false,"max_input(...TRUNCATED) | 0.833748 | 6 | 9 | "Create a new change request with a value of \"zootomical collumelliaceous saleratus enolizable Rudo(...TRUNCATED) | null | null | 0 | 475 | 1 | 4 | train | {"agent_name":"GenericAgent-moonshotai_kimi-k2-0905","chat_model_args":{"log_probs":false,"max_input(...TRUNCATED) | "<s><|system|>\nYou are a thoughtful and systematic AI assistant built by ServiceNow Language Models(...TRUNCATED) | [
[
0,
20624
]
] |
KIMI | 117 | "You are an agent trying to solve a web task based on the content of the page and\nuser instructions(...TRUNCATED) | "# Instructions\nReview the current state of the page and all other information to find the best\npo(...TRUNCATED) | "\n<think>\nI can see that I've successfully clicked on the \"Closure Information\" tab and now have(...TRUNCATED) | 6,013 | workarena.servicenow.create-change-request | {"agent_name":"GenericAgent-moonshotai_kimi-k2-0905","chat_model_args":{"log_probs":false,"max_input(...TRUNCATED) | 0.833748 | 7 | 9 | "Create a new change request with a value of \"zootomical collumelliaceous saleratus enolizable Rudo(...TRUNCATED) | null | null | 0 | 475 | 1 | 4 | train | {"agent_name":"GenericAgent-moonshotai_kimi-k2-0905","chat_model_args":{"log_probs":false,"max_input(...TRUNCATED) | "<s><|system|>\nYou are a thoughtful and systematic AI assistant built by ServiceNow Language Models(...TRUNCATED) | [
[
0,
21205
]
] |
End of preview. Expand in Data Studio
README.md exists but content is empty.
- Downloads last month
- 13