Dataset Viewer
Auto-converted to Parquet Duplicate
original
stringlengths
173
40.1k
prompt
dict
response_0
stringlengths
9
35k
response_1
stringlengths
0
30.1k
generator_model
stringclasses
1 value
generation_params
stringclasses
1 value
final_response
stringlengths
59
30.1k
jaccard_1
float64
0
1
jaccard_2
float64
0
1
levenshtein
float64
0
38.1k
softngram
float64
0
1
cosdist
float64
-0.01
0.96
bertscore_precision
float64
-0
0.42
bertscore_recall
float64
-0
0.59
bertscore
float64
-0
0.5
moverscore
float64
0.02
1.13
reranker
float64
-10.56
17.1
original_perplexity_llama_instruct
float64
1.43
561
original_entropy_llama_instruct
float64
0.39
7.11
original_topp_outlier_llama_instruct
float64
0
0.15
original_topk_outlier_llama_instruct
float64
0
0.55
original_fastdetectgpt_llama_instruct
float64
-21.82
21.1
original_perplexity_llama_base
float64
1.03
378
original_entropy_llama_base
float64
0.04
5.8
original_topp_outlier_llama_base
float64
0
0.12
original_topk_outlier_llama_base
float64
0
0.52
original_fastdetectgpt_llama_base
float64
-7.27
5.34
final_response_perplexity_llama_instruct
float64
1.52
561
final_response_entropy_llama_instruct
float64
0.42
7.12
final_response_topp_outlier_llama_instruct
float64
0
0.23
final_response_topk_outlier_llama_instruct
float64
0
0.55
final_response_fastdetectgpt_llama_instruct
float64
-21.06
20.3
final_response_perplexity_llama_base
float64
1.13
374
final_response_entropy_llama_base
float64
0.17
5.79
final_response_topp_outlier_llama_base
float64
0
0.16
final_response_topk_outlier_llama_base
float64
0
0.52
final_response_fastdetectgpt_llama_base
float64
-20.51
6.91
original_binoculars
float64
0.01
1.18
final_response_binoculars
float64
0.05
1.1
original_editlens_bucket_roberta_large
int64
0
3
original_editlens_score_roberta_large
float64
0.01
0.99
final_response_editlens_bucket_roberta_large
int64
0
3
final_response_editlens_score_roberta_large
float64
0.01
1
Thank you for sharing this sad update. It began early in June. Diego would leave the litter box with a wet back end. I'd have to try to catch him and dry him off. He was also being aggressive towards Cali so after consultation with the... Had Cali at the vet for a health check and shots. Also had the vet look at the tw...
{ "chat_turns": [ "<document>\n{{DOC}}\n</document>\n\nChange the tone of this text to be more relaxed and friendly.\nOutput the full new text with no extra statements or commentations.\nBegin directly with the text itself. Do not add a title, a heading, or a label naming what you have written." ], "use_multi...
Hey there! 😊 So, a little update on our furry family. Diego started having a bit of a wet back end in early June, and I’d have to gently dry him off after he left the litter box. He was also getting a bit feisty with Cali, so we took them both to the vet for a check-up and some shots. Good news – one of Cali’s lumps ...
cyankiwi/granite-4.1-8b-AWQ-INT4
{"temperature": 0.6, "top_p": 0.9, "presence_penalty": 1.5, "extra_body": {"top_k": 40}}
Hey there! 😊 So, a little update on our furry family. Diego started having a bit of a wet back end in early June, and I’d have to gently dry him off after he left the litter box. He was also getting a bit feisty with Cali, so we took them both to the vet for a check-up and some shots. Good news – one of Cali’s lumps o...
0.730769
0.898167
1,069
0.700098
0.121611
0.154862
0.135461
0.145272
0.571387
-5.3125
29.431168
2.968705
0.052308
0.132308
-3.099721
21.108406
3.051931
0.046154
0.107692
0.017457
12.240019
2.205429
0.052632
0.075188
-3.005851
11.120249
2.555337
0.025063
0.070175
1.392136
0.912657
0.912592
0
0.009655
1
0.409237
CLXVI Sanjaya said, Karna, the son of Vikartana O king, resisted the mighty car-warrior Sahadeva in that battle, who advanced from desire of getting at Drona. Piercing the son of Radha with nine shafts, Sahadeva once more pierced that warrior with nine straight arrows. Karna then pierced Sahadeva in return with a hundr...
{ "chat_turns": [ "<document>\n{{DOC}}\n</document>\n\nModify this text to be as long as possible. At least half of the original text must be repeated verbatim, and at most three quarters of it; the remainder must be new prose.\nOutput the full new text with no extra statements or commentations.\nBegin directly w...
CLXVI Sanjaya said, Karna, the son of Vikartana O king, resisted the mighty car-warrior Sahadeva in that battle, who advanced from desire of getting at Drona. Piercing the son of Radha with nine shafts, Sahadeva once more pierced that warrior with nine straight arrows. Karna then pierced Sahadeva in return with a hundr...
cyankiwi/granite-4.1-8b-AWQ-INT4
{"temperature": 0.6, "top_p": 0.9, "presence_penalty": 1.5, "extra_body": {"top_k": 40}}
CLXVI Sanjaya said, Karna, the son of Vikartana O king, resisted the mighty car-warrior Sahadeva in that battle, who advanced from desire of getting at Drona. Piercing the son of Radha with nine shafts, Sahadeva once more pierced that warrior with nine straight arrows. Karna then pierced Sahadeva in return with a hundr...
0.272189
0.257143
1,048
0.236559
0.12986
0.042506
0.008047
0.025581
0.23648
-5.875
14.173077
2.463246
0.05227
0.115543
-2.228704
8.895827
2.265456
0.035763
0.07978
0.995518
11.902286
2.321683
0.049145
0.100427
-2.171867
8.743671
2.239163
0.035256
0.077991
1.022023
0.754622
0.789528
0
0.023015
0
0.023015
This processor is one of the most famous one in guitar racks and old studios. It was introduced in 1985 and quickly became successful, it was relatively cheap, and has some very interesting effects algorithms. One of them is the famous ‘Pitch Change C’ algorithm, used by many legendary session guitarists in that era, i...
{ "chat_turns": [ "<document>\n{{DOC}}\n</document>\n\nTake inspiration from the style, language, and content of this text, and write a new piece of text on a topic of your choice. However, you should write the new text as if it was a rough draft being typed out on a phone.\nOutput the full new text with no extra...
I recently decided to dive into the world of vintage synthesizers, specifically a classic Roland Juno-6 that I stumbled upon at a garage sale. This synth, introduced in the early '80s, is renowned for its lush analog sounds and iconic filter sweeps, making it a staple in both electronic and pop music during its heyday....
cyankiwi/granite-4.1-8b-AWQ-INT4
{"temperature": 0.6, "top_p": 0.9, "presence_penalty": 1.5, "extra_body": {"top_k": 40}}
I recently decided to dive into the world of vintage synthesizers, specifically a classic Roland Juno-6 that I stumbled upon at a garage sale. This synth, introduced in the early '80s, is renowned for its lush analog sounds and iconic filter sweeps, making it a staple in both electronic and pop music during its heyday....
0.867052
0.965714
1,539
0.905363
0.465163
0.214928
0.183499
0.199522
0.66981
3.6875
30.019443
3.32871
0.046099
0.131206
-0.487966
22.984729
3.058342
0.049645
0.113475
-0.5479
6.978194
1.800024
0.041363
0.048662
-1.603477
6.3076
1.893264
0.026764
0.041363
0.577181
0.846486
0.860985
0
0.013802
3
0.999549
This post explains the difference between foodservice operations that are outsourced to contract management companies and those that are 'self-operated' (self-op). You’re definitely going to want to sell to these folks. It’s not an either/or, self-op or contract-managed, situation. You’re going to want to go after both...
{ "chat_turns": [ "<document>\n{{DOC}}\n</document>\n\nChange the tone of this text to be more empathetic and understanding.\nOutput the full new text with no extra statements or commentations.\nBegin directly with the text itself. Do not add a title, a heading, or a label naming what you have written." ], "u...
It's truly exciting to consider the unique opportunities that both outsourced and self-operated foodservice operations present for your business. Understanding the distinctions between these two approaches can guide you in crafting a strategy that resonates deeply with each type of institution, allowing you to build me...
cyankiwi/granite-4.1-8b-AWQ-INT4
{"temperature": 0.6, "top_p": 0.9, "presence_penalty": 1.5, "extra_body": {"top_k": 40}}
It's truly exciting to consider the unique opportunities that both outsourced and self-operated foodservice operations present for your business. Understanding the distinctions between these two approaches can guide you in crafting a strategy that resonates deeply with each type of institution, allowing you to build me...
0.822857
0.969199
2,820
0.834768
0.102921
0.174497
0.181043
0.177783
0.645246
-5.375
12.161763
2.362016
0.057598
0.072304
-1.86494
9.853105
2.438624
0.033088
0.061275
2.1153
9.992292
2.276658
0.048924
0.068493
-0.292695
9.28057
2.259481
0.039139
0.054795
0.376122
0.877299
0.866542
1
0.301078
3
0.99957
Personalized baskets filled with a selection of foods, personal care products and specialty gifts, hand-crafted in Maine. Biologics, or drugs derived from living cells, often are deemed to be specialty drugs. In the Extensions tab, SDKs that apply to the project from which the Reference Manager dialog box was invoked a...
{ "chat_turns": [ "<document>\n{{DOC}}\n</document>\n\nWrite a detailed descriptor for the style of writing of this text, including its qualitative and quantitative properties.\nDo not output anything besides what you were requested to write, and do not output any extra commentary.", "{{RESP_0}}\nAbove is an ...
The style of writing in this text is characterized by a highly eclectic and disjointed mixture of topics, lacking a coherent structure or thematic continuity. Qualitatively, the text exhibits: 1. **Lack of Cohesion**: The content jumps abruptly from one subject to another without logical transitions, making it difficu...
It got pretty addicting to watch the biologics market expand, especially with the new Gunsmith system that allowed for more precise modifications. Meanwhile, I was browsing through some personal care products at the local store, wondering if the latest shampoo would really make a difference in my hair's texture. As I s...
cyankiwi/granite-4.1-8b-AWQ-INT4
{"temperature": 0.6, "top_p": 0.9, "presence_penalty": 1.5, "extra_body": {"top_k": 40}}
It got pretty addicting to watch the biologics market expand, especially with the new Gunsmith system that allowed for more precise modifications. Meanwhile, I was browsing through some personal care products at the local store, wondering if the latest shampoo would really make a difference in my hair's texture. As I s...
0.93698
0.985617
6,211
0.958229
0.756852
0.17024
0.236439
0.204715
0.70685
9.5
43.871325
3.602982
0.055921
0.221053
-2.661363
34.701647
3.389289
0.058553
0.198684
-2.479291
15.273513
2.696525
0.04878
0.126829
-0.196562
16.312344
2.974217
0.039024
0.107317
1.159068
0.902964
0.910813
0
0.015037
3
0.998489
The indie comedy "Me," recently of the Marfa Film Festival, is an always amusing and frequently hilarious commentary on reality shows. The blurred lines between reality and "reality" include writer/producer/director/star Jefery Levy, who is also the creative force behind the recently reviewed sensual drama "The Key," p...
{ "chat_turns": [ "<document>\n{{DOC}}\n</document>\n\nModify this text to be as long as possible. At least half of the original text must be repeated verbatim, and at most three quarters of it; the remainder must be new prose.\nOutput the full new text with no extra statements or commentations.\nBegin directly w...
The indie comedy "Me," recently of the Marfa Film Festival, is an always amusing and frequently hilarious commentary on reality shows. The blurred lines between reality and "reality" include writer/producer/director/star Jefery Levy, who is also the creative force behind the recently reviewed sensual drama "The Key," p...
The indie comedy "Me," recently of the Marfa Film Festival, is an always amusing and frequently hilarious commentary on reality shows. The blurred lines between reality and "reality" include writer/producer/director/star Jefery Levy, who is also the creative force behind the recently reviewed sensual drama "The Key," p...
cyankiwi/granite-4.1-8b-AWQ-INT4
{"temperature": 0.6, "top_p": 0.9, "presence_penalty": 1.5, "extra_body": {"top_k": 40}}
The indie comedy "Me," recently of the Marfa Film Festival, is an always amusing and frequently hilarious commentary on reality shows. The blurred lines between reality and "reality" include writer/producer/director/star Jefery Levy, who is also the creative force behind the recently reviewed sensual drama "The Key," p...
0.641914
0.668288
5,204
0.657768
0.045973
0.153295
0.021033
0.091955
0.522748
-4.8125
28.527505
3.43263
0.059322
0.188559
0.707636
20.662563
3.054321
0.052966
0.154661
0.237889
9.064062
2.127413
0.053857
0.080058
-1.461245
7.56869
2.261505
0.024745
0.065502
4.375905
0.799293
0.797781
0
0.053711
0
0.108552
Hello all Have only kept for maybe 9/10 games across the last couple of years, but as of next season Ill be keeping full time so am going to be after a new pair of gloves. Been using a pair of Youth Sized B3s but they feel a little too small and havent softened up at all. The problem is that whilst youth sized are a bi...
{ "chat_turns": [ "<document>\n{{DOC}}\n</document>\n\nEdit this text by replacing, removing or inserting twenty to thirty words, without changing anything else.\nOutput the full new text with no extra statements or commentations.\nBegin directly with the text itself. Do not add a title, a heading, or a label nam...
Hello all, I've only kept records for about 9 out of 10 games over the past couple of years, but starting next season, I'll be keeping full-time and need a new pair of gloves. I've been using Youth Sized B3 gloves, but they feel too small and haven't softened up. The issue is that while youth sizes are cramped, regula...
I've only kept records for about 9 out of 10 games over the past couple of years, but starting next season, I'll be keeping full-time and need a new pair of gloves. I've been using Youth Sized B3 gloves, but they feel too small and haven't softened up. The issue is that while youth sizes are cramped, regular men's size...
cyankiwi/granite-4.1-8b-AWQ-INT4
{"temperature": 0.6, "top_p": 0.9, "presence_penalty": 1.5, "extra_body": {"top_k": 40}}
I've only kept records for about 9 out of 10 games over the past couple of years, but starting next season, I'll be keeping full-time and need a new pair of gloves. I've been using Youth Sized B3 gloves, but they feel too small and haven't softened up. The issue is that while youth sizes are cramped, regular men's size...
0.664948
0.871972
429
0.332998
0.166599
0.116008
0.132431
0.124296
0.521487
-4
32.758523
2.832128
0.084906
0.141509
-4.054773
20.654534
3.042147
0.056604
0.103774
0.089157
22.311907
2.619855
0.096257
0.149733
-3.036921
20.056152
2.987575
0.048128
0.128342
-0.0662
0.907436
0.953395
0
0.008354
1
0.394476
'Stars of Science' regularly showcases inventions across multiple disciplines and sectors that help communities in the Arab world. Multiple contestants throughout the history of the show have developed inventions that focus on preserving their communities’ identity and culture. Two contestants on the current season 13 ...
{ "chat_turns": [ "<document>\n{{DOC}}\n</document>\n\nModify this text to be as long as possible. At least half of the original text must be repeated verbatim, and at most three quarters of it; the remainder must be new prose.\nOutput the full new text with no extra statements or commentations.\nBegin directly w...
'Stars of Science' regularly showcases inventions across multiple disciplines and sectors that help communities in the Arab world. Multiple contestants throughout the history of the show have developed inventions that focus on preserving their communities’ identity and culture. Two contestants on the current season 13...
cyankiwi/granite-4.1-8b-AWQ-INT4
{"temperature": 0.6, "top_p": 0.9, "presence_penalty": 1.5, "extra_body": {"top_k": 40}}
'Stars of Science' regularly showcases inventions across multiple disciplines and sectors that help communities in the Arab world. Multiple contestants throughout the history of the show have developed inventions that focus on preserving their communities’ identity and culture. Two contestants on the current season 13 ...
0.189252
0.204019
1,010
0.182551
0.007646
0.038291
0.01094
0.024807
0.232088
-3.6875
21.677741
2.968342
0.053748
0.140028
-1.231804
16.629358
2.624869
0.056577
0.117397
-2.310942
16.594063
2.796506
0.045977
0.114943
-0.164582
13.501659
2.499595
0.049425
0.104598
-1.46301
0.861881
0.841646
0
0.096332
0
0.096332
I didn t want to those services, but three jobs statewide. Book offer great places to sit at 0.6 miles 15 years, mandarin. Regulators informally vetoed a planned expansion, announced at a bath seneca model! Additional fees are working closely reviewing cases of this year on the first day, there s website. Still take ad...
{ "chat_turns": [ "<document>\n{{DOC}}\n</document>\n\nWrite a new piece of text that is on a completely different topic, but reuses verbatim as many phrases and sentences from this text as possible.\nOutput the full new text with no extra statements or commentations.\nBegin directly with the text itself. Do not ...
Mgp will remain. Typically, if you can result. Have disease control board. Mail-In-Only voting ended march 16. Shares surged to mgm national harbor resort spa salon. Ricky martin: the city casino resort and human services. Mlb and washington and four shows and the novel coronavirus. Sometimes it will be pushed some cer...
cyankiwi/granite-4.1-8b-AWQ-INT4
{"temperature": 0.6, "top_p": 0.9, "presence_penalty": 1.5, "extra_body": {"top_k": 40}}
Mgp will remain. Typically, if you can result. Have disease control board. Mail-In-Only voting ended march 16. Shares surged to mgm national harbor resort spa salon. Ricky martin: the city casino resort and human services. Mlb and washington and four shows and the novel coronavirus. Sometimes it will be pushed some cer...
0.686007
0.713137
1,616
0.025714
0.305955
0.036997
0.148212
0.096012
0.536832
-5.5625
141.439919
5.440916
0.035573
0.33004
4.03109
91.023011
4.520366
0.055336
0.284585
0.079273
217.94439
5.624142
0.039735
0.377483
1.069639
98.091889
4.485186
0.072848
0.298013
-0.471177
0.770027
0.739268
0
0.013885
0
0.016192
As a follow up from last week, Monday evening the French President Macron as expected extended our lockdown period to May 11, 2020. However, with this came discussion about how France would slowly start to reopen on that date. A phased in approach starting with getting kids back to school and other industries starting ...
{ "chat_turns": [ "<document>\n{{DOC}}\n</document>\n\nTranslate the given text to Chinese.\nDo not output anything besides what you were requested to write, and do not output any extra commentary.", "{{RESP_0}}\nTranslate this text to English. Output only the English translation with no extra commentary.\nBe...
上周的后续行动中,周一晚上,法国总统马克龙如预期般将封锁期延长至2020年5月11日。然而,随之而来的是关于法国何时逐步恢复运作的讨论。 采取分阶段方法,首先让孩子们回到学校,并让其他行业重新开放,引发了乐观情绪。从旅游/旅游角度来看,马克龙尚未解除餐馆的限制,非欧洲人(不包括英国)进入法国进行非必要旅行的国际边境仍然关闭。 这意味着什么? 目前还没有人能确定。我们知道两家英国航空公司已经建议,非必要旅行的航班将于6月中旬重新启动。欧盟委员会建议在欧洲内部恢复旅行,这对于仍希望今年与我们旅行的英国和欧洲客户来说是好消息。更令人振奋的消息,对于非欧洲客户在秋季有一丝希望——即《困难之路》(Tour de France)将于2020...
On Monday night, French President Macron as expected extended the lockdown until May 11, 2020. However, this has sparked discussions about when France will gradually resume operations. Taking a phased approach, first allowing children to return to school and then reopening other sectors has generated optimistic sentim...
cyankiwi/granite-4.1-8b-AWQ-INT4
{"temperature": 0.6, "top_p": 0.9, "presence_penalty": 1.5, "extra_body": {"top_k": 40}}
On Monday night, French President Macron as expected extended the lockdown until May 11, 2020. However, this has sparked discussions about when France will gradually resume operations. Taking a phased approach, first allowing children to return to school and then reopening other sectors has generated optimistic sentime...
0.578261
0.790026
607
0.373093
0.040086
0.101076
0.107215
0.104156
0.469829
-8.8125
23.144369
2.90636
0.05102
0.112245
-1.771961
17.595691
2.809396
0.044218
0.091837
-0.454493
22.793286
2.897544
0.049618
0.137405
-1.652082
18.635031
2.765109
0.034351
0.114504
-1.198791
0.886347
0.911059
0
0.010777
1
0.292401
When I committed to NaBloPoMo, I’m not sure that I realized the commitment it would entail. Well, sure, I knew that it meant that I would post daily. It meant that I would be able to refresh my blog with a multitude of recipes that I’ve been dying to share, but for one reason or another, hadn’t blogged yet. The main re...
{ "chat_turns": [ "<document>\n{{DOC}}\n</document>\n\nWrite a detailed descriptor for the manner in which the author of this uses language, and things such as their word choice.\nDo not output anything besides what you were requested to write, and do not output any extra commentary.", "{{RESP_0}}\nAbove is a...
The author employs a warm, inviting, and descriptive language style, characterized by the following elements: 1. **Personal Tone**: The author uses a first-person narrative, sharing personal reflections and motivations behind committing to NaBloPoMo, which creates an intimate connection with the reader. 2. **Reflecti...
I’m not sure that I realized the commitment it would entail when I decided to take on NaBloPoMo, but here I am, ready to dive in. Life is busy, and sometimes it’s more convenient to write on a Sunday rather than during the hectic weekdays. Today, I want to share something personal—my journey with blogging and a little ...
cyankiwi/granite-4.1-8b-AWQ-INT4
{"temperature": 0.6, "top_p": 0.9, "presence_penalty": 1.5, "extra_body": {"top_k": 40}}
I’m not sure that I realized the commitment it would entail when I decided to take on NaBloPoMo, but here I am, ready to dive in. Life is busy, and sometimes it’s more convenient to write on a Sunday rather than during the hectic weekdays. Today, I want to share something personal—my journey with blogging and a little ...
0.801115
0.944561
2,258
0.827208
0.243002
0.187242
0.196023
0.191656
0.671135
1.9375
7.319979
1.787634
0.061058
0.050204
-2.901436
4.979841
1.663907
0.027137
0.036635
0.925935
4.510257
1.392892
0.045741
0.029968
-1.884659
4.22261
1.600844
0.025237
0.026814
2.485772
0.718811
0.807867
0
0.040475
3
0.733539
- Bright Star reviews, ratings etc. *SPOILERS* by Saturn » Sun May 17, 2009 2:56 pm - 312 Replies - 1189380 Views - Last post by Pjerrot Sun Jun 06, 2010 11:38 pm - - Bright Star the movie website! by Saturn » Thu Jun 12, 2008 9:11 pm - 47 Replies - 185336 Views - Last post by Raphael Thu Nov 26, 2009 3:27 pm - - Discu...
{ "chat_turns": [ "<document>\n{{DOC}}\n</document>\n\nEdit this text by replacing, removing or inserting twenty to thirty words, without changing anything else.\nOutput the full new text with no extra statements or commentations.\nBegin directly with the text itself. Do not add a title, a heading, or a label nam...
Bright Star reviews, ratings etc. *SPOILERS* by Saturn » Sun May 17, 2009 2:56 pm - 312 Replies - 1189380 Views - Last post by Pjerrot Sun Jun 06, 2010 11:38 pm Bright Star the movie website! by Saturn » Thu Jun 12, 2008 9:11 pm - 47 Replies - 185336 Views - Last post by Raphael Thu Nov 26, 2009 ...
cyankiwi/granite-4.1-8b-AWQ-INT4
{"temperature": 0.6, "top_p": 0.9, "presence_penalty": 1.5, "extra_body": {"top_k": 40}}
Bright Star reviews, ratings etc. *SPOILERS* by Saturn » Sun May 17, 2009 2:56 pm - 312 Replies - 1189380 Views - Last post by Pjerrot Sun Jun 06, 2010 11:38 pm Bright Star the movie website! by Saturn » Thu Jun 12, 2008 9:11 pm - 47 Replies - 185336 Views - Last post by Raphael Thu Nov 26, 2009 3:27 pm Discussion abou...
0.022857
0.093278
162
0.000119
0.006475
0.020975
0.018915
0.019946
0.22074
-7.8125
4.211267
1.436177
0.023256
0.06538
-0.057009
3.770522
1.358182
0.014919
0.057043
1.15384
4.372434
1.463621
0.025397
0.066667
-0.420288
3.916807
1.396796
0.014512
0.057596
1.143624
0.84123
0.849453
0
0.037114
0
0.03237
CAPSULE: Kaiju Noir. With the Japanese franchise on the King of the Monsters in hiatus, the Godzilla character is being loaned to Warner Brothers so that Gareth Edwards can make an American Godzilla film. This is a script whose drama is better than Toho's usual fare, but audiences may find the new film is dark and drab...
{ "chat_turns": [ "<document>\n{{DOC}}\n</document>\n\nEdit this text by completely rewriting the topic sentence of every paragraph, while preserving the supporting sentences in those paragraphs verbatim.\nOutput the full new text with no extra statements or commentations.\nBegin directly with the text itself. Do...
The new American Godzilla film, directed by Gareth Edwards and starring Bryan Cranston, marks a significant departure from Toho's traditional kaiju fare, offering a darker, more somber narrative that revisits the original film's themes of tragedy and nuclear consequence. In this version, the opening scene shifts focus ...
In this new American Godzilla film, directed by Gareth Edwards and starring Bryan Cranston, a significant departure from Toho's traditional kaiju fare is evident, offering a darker, more somber narrative that revisits the original film's themes of tragedy and nuclear consequence, as the opening scene shifts focus to a ...
cyankiwi/granite-4.1-8b-AWQ-INT4
{"temperature": 0.6, "top_p": 0.9, "presence_penalty": 1.5, "extra_body": {"top_k": 40}}
In this new American Godzilla film, directed by Gareth Edwards and starring Bryan Cranston, a significant departure from Toho's traditional kaiju fare is evident, offering a darker, more somber narrative that revisits the original film's themes of tragedy and nuclear consequence, as the opening scene shifts focus to a ...
0.872564
0.97343
4,541
0.895899
0.283138
0.161746
0.187128
0.174632
0.651956
-6.0625
20.046873
3.04344
0.05098
0.131765
0.670267
15.815705
2.680367
0.058824
0.111373
-1.282028
15.370043
2.726381
0.044496
0.11007
-0.056481
14.011865
2.664633
0.030445
0.100703
0.2361
0.803702
0.848266
0
0.023796
3
0.988863
Bol Entertainment is an upcoming Pakistani channel that is a subsidiary project of Axact organization, which is the world’s fastest leading Karachi-based IT company. Pakistan Electronic Media Regulatory Authority (PEMRA) has approved radio and satellite television licenses to the BOL Network. The founders of the channe...
{ "chat_turns": [ "<document>\n{{DOC}}\n</document>\n\nChange the tone of this text to be more relaxed and friendly.\nOutput the full new text with no extra statements or commentations.\nBegin directly with the text itself. Do not add a title, a heading, or a label naming what you have written.", "{{RESP_0}}\...
Hey there! So, you might have heard about Bol Entertainment, the super cool new Pakistani channel that's part of the awesome Axact organization. Axact is this super-fast IT company based in Karachi, and they're really shaking things up. The Pakistan Electronic Media Regulatory Authority (PEMRA) just gave the thumbs-up ...
BOL Entertainment, the innovative new Pakistani channel under the umbrella of the dynamic Axact organization, is indeed generating a lot of excitement. Axact, a rapidly growing IT company headquartered in Karachi, is at the forefront of transforming digital landscapes. With the recent approval from the Pakistan Electro...
cyankiwi/granite-4.1-8b-AWQ-INT4
{"temperature": 0.6, "top_p": 0.9, "presence_penalty": 1.5, "extra_body": {"top_k": 40}}
BOL Entertainment, the innovative new Pakistani channel under the umbrella of the dynamic Axact organization, is indeed generating a lot of excitement. Axact, a rapidly growing IT company headquartered in Karachi, is at the forefront of transforming digital landscapes. With the recent approval from the Pakistan Electro...
0.848889
0.946328
1,656
0.869361
0.191256
0.240841
0.125479
0.187233
0.719341
2.375
28.105714
2.704589
0.063158
0.168421
-2.611264
22.5573
2.542473
0.115789
0.136842
-2.523425
7.200448
1.964177
0.032698
0.046322
-0.103899
6.38383
2.012511
0.021798
0.049046
1.661662
1.019556
0.807326
1
0.25899
3
0.999501
SAT Practice: Essay - Amazon Web Services. Scores for the samples provided below were assigned on a 1-4 scale according to the redesigned SAT Essay Scoring Rubric. It is important to note that although these are representative samples of student ability at each score point, the set itself does not exhaustively illustra...
{ "chat_turns": [ "<document>\n{{DOC}}\n</document>\n\nChange the tone of this text to be more suited for a business audience.\nOutput the full new text with no extra statements or commentations.\nBegin directly with the text itself. Do not add a title, a heading, or a label naming what you have written." ], ...
SAT Practice: Essay - Amazon Web Services. Scores for the samples provided below were assigned on a 1-4 scale according to the redesigned SAT Essay Scoring Rubric. These samples represent student ability at each score point but do not exhaustively cover the range of skills in Reading, Analysis, and Writing associated ...
cyankiwi/granite-4.1-8b-AWQ-INT4
{"temperature": 0.6, "top_p": 0.9, "presence_penalty": 1.5, "extra_body": {"top_k": 40}}
SAT Practice: Essay - Amazon Web Services. Scores for the samples provided below were assigned on a 1-4 scale according to the redesigned SAT Essay Scoring Rubric. These samples represent student ability at each score point but do not exhaustively cover the range of skills in Reading, Analysis, and Writing associated w...
0.717557
0.864211
2,662
0.312343
0.042347
0.098328
0.139089
0.11918
0.528709
-7.4375
15.777492
2.535348
0.065817
0.096603
-3.07108
8.389224
2.110075
0.044586
0.069002
-0.254875
19.837715
2.816855
0.05
0.114
-1.663618
19.182255
2.544328
0.068
0.12
-4.191767
0.608575
0.904538
0
0.008327
1
0.283292
End of preview. Expand in Data Studio

Auto-Generated FastDetector Dataset

  • Dataset: G-reen/cc-2021-stat
  • Globals Config: config/globals.toml
  • Analysis Config: config/analysis_nofilter.toml
  • Rows: 23,214

Evaluation Results

  • Prompt Subsets: 4 (direct_reference, indirect_reference, revise, rewrite)
  • Generator Configs: 7 (Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 0.6), Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 1.25), Ministral-3-8B-Instruct-2512-AWQ-4bit (Temp: 0.6), Qwen3-8B-AWQ (Temp: 0.7), gemma-4-E4B-it (Temp: 0.7), granite-4.1-8b-AWQ-INT4 (Temp: 0.6), granite-4.1-8b-AWQ-INT4 (Temp: 1.25))
  • Classifiers: 13 (EditLens Roberta-Large Score, EditLens Roberta-Large Bucket, Perplexity (Llama-3.2-3B-Instruct), Perplexity (Llama-3.2-3B), Entropy (Llama-3.2-3B-Instruct), Entropy (Llama-3.2-3B), Top-p Outliers (Llama-3.2-3B-Instruct), Top-p Outliers (Llama-3.2-3B), Top-k Outliers (Llama-3.2-3B-Instruct), Top-k Outliers (Llama-3.2-3B), FastDetectGPT (Llama-3.2-3B-Instruct), FastDetectGPT (Llama-3.2-3B), Binoculars)
  • Filter Conditions: None
  • Evaluation / Validation Rows: 20,892 / 2,322 (validation_size = 0.1)
  • Base Columns: original (Human), final_response (AI)

The best classifier was EditLens Roberta-Large Score with an AUROC of 0.8775. The hardest prompt subset was rewrite with a TPR of 0.0536, and the hardest generator config was gemma-4-E4B-it (Temp: 0.7) with a TPR of 0.0693.

Classifier Threshold AUROC TPR FPR Accuracy F1
✔️ EditLens Roberta-Large Score 0.5961 0.8775 0.3588 0.0046 0.6771 0.5263
EditLens Roberta-Large Bucket 0.0000 0.8145 0.6656 0.0513 0.8072 0.7754
Binoculars 1.0472 0.6302 0.0121 0.0031 0.5045 0.0239
FastDetectGPT (Llama-3.2-3B-Instruct) 6.7770 0.5725 0.0027 0.0041 0.4993 0.0054
Top-k Outliers (Llama-3.2-3B-Instruct) 0.0109 0.5637 0.0191 0.0033 0.5079 0.0373
Top-p Outliers (Llama-3.2-3B-Instruct) 0.0158 0.5598 0.0108 0.0030 0.5039 0.0213
Perplexity (Llama-3.2-3B-Instruct) 2.2698 0.5478 0.0055 0.0045 0.5005 0.0108
Entropy (Llama-3.2-3B-Instruct) 0.7575 0.5391 0.0032 0.0044 0.4994 0.0063
Top-p Outliers (Llama-3.2-3B) 0.0054 0.5168 0.0062 0.0071 0.4995 0.0122
Top-k Outliers (Llama-3.2-3B) 0.0056 0.5168 0.0053 0.0048 0.5002 0.0104
FastDetectGPT (Llama-3.2-3B) -3.2326 0.5159 0.0986 0.0066 0.5460 0.1784
Entropy (Llama-3.2-3B) 0.2888 0.4869 0.0005 0.0049 0.4978 0.0010
❗ Perplexity (Llama-3.2-3B) 1.2916 0.4854 0.0005 0.0044 0.4980 0.0010

Classifier metrics averaged within each prompt and generator subset:

Subset Average AUROC Average TPR Average FPR Average Accuracy Average F1
✔️ Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 0.6) 0.6594 0.0700 0.0074 0.5313 0.1065
Model: Qwen3-8B-AWQ (Temp: 0.7) 0.6285 0.0704 0.0077 0.5313 0.1015
Model: Ministral-3-8B-Instruct-2512-AWQ-4bit (Temp: 0.6) 0.6228 0.1264 0.0086 0.5589 0.1383
Prompt: revise 0.6184 0.1120 0.0082 0.5519 0.1353
Model: granite-4.1-8b-AWQ-INT4 (Temp: 0.6) 0.6115 0.0940 0.0086 0.5427 0.1174
Prompt: indirect_reference 0.6009 0.1086 0.0076 0.5505 0.1445
Prompt: direct_reference 0.5961 0.0921 0.0086 0.5417 0.1263
Model: granite-4.1-8b-AWQ-INT4 (Temp: 1.25) 0.5726 0.1051 0.0083 0.5484 0.1302
Model: gemma-4-E4B-it (Temp: 0.7) 0.5472 0.0693 0.0069 0.5312 0.0935
Prompt: rewrite 0.5314 0.0536 0.0081 0.5227 0.0812
❗ Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 1.25) 0.4643 0.1065 0.0097 0.5484 0.1412

✔️ marks the best AUROC, ❗ the worst.

Statistics of Interest

JACCARD_1 by Prompt Subset COSDIST by Prompt Subset JACCARD_1 by Generator Config COSDIST by Generator Config

Appendix

Table of contents

  1. Univariate Analysis
  2. Correlation Heatmap
  3. Distance Histograms
  4. Distance Histograms per Prompt Subset
  5. Distance Histograms per Generator Config Subset
  6. Classifier: EditLens Roberta-Large Score
  7. Classifier: EditLens Roberta-Large Bucket
  8. Classifier: Perplexity (Llama-3.2-3B-Instruct)
  9. Classifier: Perplexity (Llama-3.2-3B)
  10. Classifier: Entropy (Llama-3.2-3B-Instruct)
  11. Classifier: Entropy (Llama-3.2-3B)
  12. Classifier: Top-p Outliers (Llama-3.2-3B-Instruct)
  13. Classifier: Top-p Outliers (Llama-3.2-3B)
  14. Classifier: Top-k Outliers (Llama-3.2-3B-Instruct)
  15. Classifier: Top-k Outliers (Llama-3.2-3B)
  16. Classifier: FastDetectGPT (Llama-3.2-3B-Instruct)
  17. Classifier: FastDetectGPT (Llama-3.2-3B)
  18. Classifier: Binoculars

Univariate Analysis

Every statistic the report does arithmetic on, over the 20,892-row evaluation split. Invalid counts rows whose value is missing or non-finite; those rows are excluded from the other columns.

Statistic N Mean Median Std Min Max Invalid
jaccard_1 20,892 0.6416 0.7434 0.2732 0.0000 1.0000 0
jaccard_2 20,892 0.7582 0.9006 0.2929 0.0000 1.0000 0
levenshtein 20,892 2308.9464 1404.0000 3256.2544 0.0000 87016.0000 0
softngram 20,892 0.5914 0.6754 0.3489 0.0000 1.0000 0
cosdist 20,892 0.2048 0.1378 0.1991 -0.0076 1.0211 0
bertscore 20,892 0.1455 0.1513 0.0808 -0.0000 0.4992 0
bertscore_precision 20,892 0.1439 0.1498 0.0825 -0.0000 0.4512 0
bertscore_recall 20,892 0.1459 0.1499 0.0842 -0.0000 0.5874 0
moverscore 20,892 0.5516 0.5975 0.2048 0.0106 1.1348 0
reranker 20,892 -2.8360 -4.6875 5.6218 -11.0000 18.7500 0
EditLens Roberta-Large Score (Human) 20,892 0.0565 0.0212 0.0915 0.0063 0.9995 0
EditLens Roberta-Large Score (AI) 20,892 0.4735 0.4029 0.3639 0.0065 0.9996 0
EditLens Roberta-Large Bucket (Human) 20,892 0.0605 0.0000 0.2908 0.0000 3.0000 0
EditLens Roberta-Large Bucket (AI) 20,892 1.3514 1.0000 1.2534 0.0000 3.0000 0
Perplexity (Llama-3.2-3B-Instruct) (Human) 20,892 18.9930 14.4788 31.4486 1.0554 2279.9307 0
Perplexity (Llama-3.2-3B-Instruct) (AI) 20,892 82.5218 13.0928 437.0904 1.0655 10003.4067 0
Perplexity (Llama-3.2-3B) (Human) 20,892 13.7984 11.0648 19.7831 1.0282 1479.0899 0
Perplexity (Llama-3.2-3B) (AI) 20,892 89.1387 11.1958 520.1588 1.0638 13124.9201 0
Entropy (Llama-3.2-3B-Instruct) (Human) 20,892 2.5271 2.4894 0.6300 0.0629 7.8258 0
Entropy (Llama-3.2-3B-Instruct) (AI) 20,892 2.6020 2.4018 1.1649 0.1077 9.3056 0
Entropy (Llama-3.2-3B) (Human) 20,892 2.3825 2.4005 0.5707 0.0411 6.3306 0
Entropy (Llama-3.2-3B) (AI) 20,892 2.5570 2.4012 0.9664 0.0909 8.3789 0
Top-p Outliers (Llama-3.2-3B-Instruct) (Human) 20,892 0.0554 0.0535 0.0163 0.0000 0.2857 0
Top-p Outliers (Llama-3.2-3B-Instruct) (AI) 20,892 0.0522 0.0507 0.0173 0.0000 0.2308 0
Top-p Outliers (Llama-3.2-3B) (Human) 20,892 0.0415 0.0414 0.0127 0.0000 0.1774 0
Top-p Outliers (Llama-3.2-3B) (AI) 20,892 0.0425 0.0401 0.0196 0.0000 0.2222 0
Top-k Outliers (Llama-3.2-3B-Instruct) (Human) 20,892 0.1060 0.0973 0.0523 0.0000 0.6452 0
Top-k Outliers (Llama-3.2-3B-Instruct) (AI) 20,892 0.1169 0.0865 0.1238 0.0000 0.8796 0
Top-k Outliers (Llama-3.2-3B) (Human) 20,892 0.0868 0.0788 0.0474 0.0000 0.5529 0
Top-k Outliers (Llama-3.2-3B) (AI) 20,892 0.1076 0.0760 0.1235 0.0000 0.8802 0
FastDetectGPT (Llama-3.2-3B-Instruct) (Human) 20,892 -1.6739 -1.7396 1.9618 -21.8239 25.2161 0
FastDetectGPT (Llama-3.2-3B-Instruct) (AI) 20,892 -1.1973 -1.3517 1.9641 -21.0610 20.2645 0
FastDetectGPT (Llama-3.2-3B) (Human) 20,892 -0.1766 -0.1449 1.1141 -7.3730 9.3717 0
FastDetectGPT (Llama-3.2-3B) (AI) 20,892 -0.7282 -0.2523 3.8187 -38.9644 16.2603 0
Binoculars (Human) 20,892 0.8280 0.8516 0.1309 0.0075 1.2099 0
Binoculars (AI) 20,892 0.8748 0.8787 0.0876 0.0257 1.1761 0

Correlation Heatmap

Pearson correlation between every statistic of interest, computed over the rows where both statistics are present.

CORRELATIONS

Distance Histograms

Distance: jaccard_1 Distance: jaccard_2 Distance: levenshtein Distance: softngram Distance: cosdist Distance: bertscore Distance: bertscore_precision Distance: bertscore_recall Distance: moverscore Distance: reranker

Distance Histograms per Prompt Subset

jaccard_1 by Prompt Subset jaccard_2 by Prompt Subset levenshtein by Prompt Subset softngram by Prompt Subset cosdist by Prompt Subset bertscore by Prompt Subset bertscore_precision by Prompt Subset bertscore_recall by Prompt Subset moverscore by Prompt Subset reranker by Prompt Subset

Distance Histograms per Generator Config Subset

jaccard_1 by Generator Config jaccard_2 by Generator Config levenshtein by Generator Config softngram by Generator Config cosdist by Generator Config bertscore by Generator Config bertscore_precision by Generator Config bertscore_recall by Generator Config moverscore by Generator Config reranker by Generator Config

Classifier: EditLens Roberta-Large Score

Performance:

Subset N AUROC TPR FPR Accuracy F1
Overall 41,784 0.8775 0.3588 0.0046 0.6771 0.5263
Prompt: direct_reference 10,356 0.8405 0.4181 0.0041 0.7070 0.5880
Prompt: indirect_reference 10,182 0.9333 0.4331 0.0055 0.7138 0.6021
Prompt: revise 10,666 0.9662 0.4568 0.0039 0.7264 0.6254
❗ Prompt: rewrite 10,580 0.7711 0.1304 0.0049 0.5628 0.2298
Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 0.6) 6,064 0.8177 0.2642 0.0040 0.6301 0.4166
Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 1.25) 6,004 0.8537 0.1666 0.0070 0.5798 0.2838
✔️ Model: Ministral-3-8B-Instruct-2512-AWQ-4bit (Temp: 0.6) 5,860 0.9907 0.6451 0.0034 0.8208 0.7826
Model: Qwen3-8B-AWQ (Temp: 0.7) 6,118 0.8077 0.2958 0.0072 0.6443 0.4541
Model: gemma-4-E4B-it (Temp: 0.7) 5,972 0.8523 0.2515 0.0027 0.6244 0.4011
Model: granite-4.1-8b-AWQ-INT4 (Temp: 0.6) 5,876 0.9029 0.4364 0.0024 0.7170 0.6066
Model: granite-4.1-8b-AWQ-INT4 (Temp: 1.25) 5,890 0.9249 0.4642 0.0054 0.7294 0.6317

Thresholding:

  • Direction: higher_is_ai
  • Swept for fpr_0_5pct with a found threshold of 0.5961.

Threshold Sweep: EditLens Roberta-Large Score

Classification Histograms:

Classifier: EditLens Roberta-Large Score

Per Prompt Subset

EditLens Roberta-Large Score: Prompt: direct_reference EditLens Roberta-Large Score: Prompt: indirect_reference EditLens Roberta-Large Score: Prompt: revise EditLens Roberta-Large Score: Prompt: rewrite

Per Generator Config Subset

EditLens Roberta-Large Score: Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 0.6) EditLens Roberta-Large Score: Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 1.25) EditLens Roberta-Large Score: Model: Ministral-3-8B-Instruct-2512-AWQ-4bit (Temp: 0.6) EditLens Roberta-Large Score: Model: Qwen3-8B-AWQ (Temp: 0.7) EditLens Roberta-Large Score: Model: gemma-4-E4B-it (Temp: 0.7) EditLens Roberta-Large Score: Model: granite-4.1-8b-AWQ-INT4 (Temp: 0.6) EditLens Roberta-Large Score: Model: granite-4.1-8b-AWQ-INT4 (Temp: 1.25)

Classifier: EditLens Roberta-Large Bucket

Performance:

Subset N AUROC TPR FPR Accuracy F1
Overall 41,784 0.8145 0.6656 0.0513 0.8072 0.7754
Prompt: direct_reference 10,356 0.7897 0.6124 0.0508 0.7808 0.7364
Prompt: indirect_reference 10,182 0.8516 0.7362 0.0517 0.8423 0.8236
Prompt: revise 10,666 0.9239 0.8785 0.0491 0.9147 0.9115
❗ Prompt: rewrite 10,580 0.6928 0.4352 0.0537 0.6907 0.5846
Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 0.6) 6,064 0.7088 0.4571 0.0508 0.7032 0.6063
Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 1.25) 6,004 0.7709 0.5936 0.0566 0.7685 0.7194
✔️ Model: Ministral-3-8B-Instruct-2512-AWQ-4bit (Temp: 0.6) 5,860 0.9743 0.9703 0.0495 0.9604 0.9608
Model: Qwen3-8B-AWQ (Temp: 0.7) 6,118 0.7343 0.5093 0.0517 0.7288 0.6526
Model: gemma-4-E4B-it (Temp: 0.7) 5,972 0.7853 0.6102 0.0492 0.7805 0.7354
Model: granite-4.1-8b-AWQ-INT4 (Temp: 0.6) 5,876 0.8527 0.7372 0.0524 0.8424 0.8239
Model: granite-4.1-8b-AWQ-INT4 (Temp: 1.25) 5,890 0.8838 0.7976 0.0489 0.8744 0.8639

Thresholding:

  • Direction: higher_is_ai
  • Swept for f1 with a found threshold of 0.0000.

Threshold Sweep: EditLens Roberta-Large Bucket

Classification Histograms:

Classifier: EditLens Roberta-Large Bucket

Per Prompt Subset

EditLens Roberta-Large Bucket: Prompt: direct_reference EditLens Roberta-Large Bucket: Prompt: indirect_reference EditLens Roberta-Large Bucket: Prompt: revise EditLens Roberta-Large Bucket: Prompt: rewrite

Per Generator Config Subset

EditLens Roberta-Large Bucket: Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 0.6) EditLens Roberta-Large Bucket: Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 1.25) EditLens Roberta-Large Bucket: Model: Ministral-3-8B-Instruct-2512-AWQ-4bit (Temp: 0.6) EditLens Roberta-Large Bucket: Model: Qwen3-8B-AWQ (Temp: 0.7) EditLens Roberta-Large Bucket: Model: gemma-4-E4B-it (Temp: 0.7) EditLens Roberta-Large Bucket: Model: granite-4.1-8b-AWQ-INT4 (Temp: 0.6) EditLens Roberta-Large Bucket: Model: granite-4.1-8b-AWQ-INT4 (Temp: 1.25)

Classifier: Perplexity (Llama-3.2-3B-Instruct)

Performance:

Subset N AUROC TPR FPR Accuracy F1
Overall 41,784 0.5478 0.0055 0.0045 0.5005 0.0108
Prompt: direct_reference 10,356 0.5778 0.0062 0.0052 0.5005 0.0122
Prompt: indirect_reference 10,182 0.5590 0.0132 0.0027 0.5052 0.0259
Prompt: revise 10,666 0.5867 0.0013 0.0051 0.4981 0.0026
Prompt: rewrite 10,580 0.4672 0.0015 0.0047 0.4984 0.0030
✔️ Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 0.6) 6,064 0.7393 0.0254 0.0030 0.5112 0.0494
❗ Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 1.25) 6,004 0.2114 0.0007 0.0057 0.4975 0.0013
Model: Ministral-3-8B-Instruct-2512-AWQ-4bit (Temp: 0.6) 5,860 0.6083 0.0007 0.0065 0.4971 0.0014
Model: Qwen3-8B-AWQ (Temp: 0.7) 6,118 0.6866 0.0046 0.0033 0.5007 0.0091
Model: gemma-4-E4B-it (Temp: 0.7) 5,972 0.4750 0.0020 0.0033 0.4993 0.0040
Model: granite-4.1-8b-AWQ-INT4 (Temp: 0.6) 5,876 0.6144 0.0037 0.0051 0.4993 0.0074
Model: granite-4.1-8b-AWQ-INT4 (Temp: 1.25) 5,890 0.4990 0.0007 0.0044 0.4981 0.0014

Thresholding:

  • Direction: lower_is_ai
  • Swept for fpr_0_5pct with a found threshold of 2.2698.

Threshold Sweep: Perplexity (Llama-3.2-3B-Instruct)

Classification Histograms:

Classifier: Perplexity (Llama-3.2-3B-Instruct)

Per Prompt Subset

Perplexity (Llama-3.2-3B-Instruct): Prompt: direct_reference Perplexity (Llama-3.2-3B-Instruct): Prompt: indirect_reference Perplexity (Llama-3.2-3B-Instruct): Prompt: revise Perplexity (Llama-3.2-3B-Instruct): Prompt: rewrite

Per Generator Config Subset

Perplexity (Llama-3.2-3B-Instruct): Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 0.6) Perplexity (Llama-3.2-3B-Instruct): Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 1.25) Perplexity (Llama-3.2-3B-Instruct): Model: Ministral-3-8B-Instruct-2512-AWQ-4bit (Temp: 0.6) Perplexity (Llama-3.2-3B-Instruct): Model: Qwen3-8B-AWQ (Temp: 0.7) Perplexity (Llama-3.2-3B-Instruct): Model: gemma-4-E4B-it (Temp: 0.7) Perplexity (Llama-3.2-3B-Instruct): Model: granite-4.1-8b-AWQ-INT4 (Temp: 0.6) Perplexity (Llama-3.2-3B-Instruct): Model: granite-4.1-8b-AWQ-INT4 (Temp: 1.25)

Classifier: Perplexity (Llama-3.2-3B)

Performance:

Subset N AUROC TPR FPR Accuracy F1
Overall 41,784 0.4854 0.0005 0.0044 0.4980 0.0010
Prompt: direct_reference 10,356 0.5068 0.0004 0.0058 0.4973 0.0008
Prompt: indirect_reference 10,182 0.4852 0.0010 0.0026 0.4992 0.0020
Prompt: revise 10,666 0.5149 0.0000 0.0049 0.4976 0.0000
Prompt: rewrite 10,580 0.4338 0.0006 0.0043 0.4981 0.0011
✔️ Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 0.6) 6,064 0.6774 0.0007 0.0023 0.4992 0.0013
❗ Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 1.25) 6,004 0.1564 0.0003 0.0063 0.4970 0.0007
Model: Ministral-3-8B-Instruct-2512-AWQ-4bit (Temp: 0.6) 5,860 0.5326 0.0000 0.0055 0.4973 0.0000
Model: Qwen3-8B-AWQ (Temp: 0.7) 6,118 0.6481 0.0007 0.0029 0.4989 0.0013
Model: gemma-4-E4B-it (Temp: 0.7) 5,972 0.4061 0.0007 0.0023 0.4992 0.0013
Model: granite-4.1-8b-AWQ-INT4 (Temp: 0.6) 5,876 0.5512 0.0007 0.0054 0.4976 0.0014
Model: granite-4.1-8b-AWQ-INT4 (Temp: 1.25) 5,890 0.4266 0.0003 0.0061 0.4971 0.0007

Thresholding:

  • Direction: lower_is_ai
  • Swept for fpr_0_5pct with a found threshold of 1.2916.

Threshold Sweep: Perplexity (Llama-3.2-3B)

Classification Histograms:

Classifier: Perplexity (Llama-3.2-3B)

Per Prompt Subset

Perplexity (Llama-3.2-3B): Prompt: direct_reference Perplexity (Llama-3.2-3B): Prompt: indirect_reference Perplexity (Llama-3.2-3B): Prompt: revise Perplexity (Llama-3.2-3B): Prompt: rewrite

Per Generator Config Subset

Perplexity (Llama-3.2-3B): Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 0.6) Perplexity (Llama-3.2-3B): Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 1.25) Perplexity (Llama-3.2-3B): Model: Ministral-3-8B-Instruct-2512-AWQ-4bit (Temp: 0.6) Perplexity (Llama-3.2-3B): Model: Qwen3-8B-AWQ (Temp: 0.7) Perplexity (Llama-3.2-3B): Model: gemma-4-E4B-it (Temp: 0.7) Perplexity (Llama-3.2-3B): Model: granite-4.1-8b-AWQ-INT4 (Temp: 0.6) Perplexity (Llama-3.2-3B): Model: granite-4.1-8b-AWQ-INT4 (Temp: 1.25)

Classifier: Entropy (Llama-3.2-3B-Instruct)

Performance:

Subset N AUROC TPR FPR Accuracy F1
Overall 41,784 0.5391 0.0032 0.0044 0.4994 0.0063
Prompt: direct_reference 10,356 0.5668 0.0029 0.0050 0.4989 0.0057
Prompt: indirect_reference 10,182 0.5409 0.0077 0.0026 0.5026 0.0152
Prompt: revise 10,666 0.5799 0.0009 0.0047 0.4981 0.0019
Prompt: rewrite 10,580 0.4678 0.0013 0.0051 0.4981 0.0026
✔️ Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 0.6) 6,064 0.7033 0.0119 0.0026 0.5046 0.0234
❗ Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 1.25) 6,004 0.2467 0.0003 0.0053 0.4975 0.0007
Model: Ministral-3-8B-Instruct-2512-AWQ-4bit (Temp: 0.6) 5,860 0.5743 0.0007 0.0058 0.4974 0.0014
Model: Qwen3-8B-AWQ (Temp: 0.7) 6,118 0.6652 0.0039 0.0033 0.5003 0.0078
Model: gemma-4-E4B-it (Temp: 0.7) 5,972 0.4608 0.0017 0.0033 0.4992 0.0033
Model: granite-4.1-8b-AWQ-INT4 (Temp: 0.6) 5,876 0.6065 0.0024 0.0051 0.4986 0.0047
Model: granite-4.1-8b-AWQ-INT4 (Temp: 1.25) 5,890 0.5167 0.0010 0.0051 0.4980 0.0020

Thresholding:

  • Direction: lower_is_ai
  • Swept for fpr_0_5pct with a found threshold of 0.7575.

Threshold Sweep: Entropy (Llama-3.2-3B-Instruct)

Classification Histograms:

Classifier: Entropy (Llama-3.2-3B-Instruct)

Per Prompt Subset

Entropy (Llama-3.2-3B-Instruct): Prompt: direct_reference Entropy (Llama-3.2-3B-Instruct): Prompt: indirect_reference Entropy (Llama-3.2-3B-Instruct): Prompt: revise Entropy (Llama-3.2-3B-Instruct): Prompt: rewrite

Per Generator Config Subset

Entropy (Llama-3.2-3B-Instruct): Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 0.6) Entropy (Llama-3.2-3B-Instruct): Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 1.25) Entropy (Llama-3.2-3B-Instruct): Model: Ministral-3-8B-Instruct-2512-AWQ-4bit (Temp: 0.6) Entropy (Llama-3.2-3B-Instruct): Model: Qwen3-8B-AWQ (Temp: 0.7) Entropy (Llama-3.2-3B-Instruct): Model: gemma-4-E4B-it (Temp: 0.7) Entropy (Llama-3.2-3B-Instruct): Model: granite-4.1-8b-AWQ-INT4 (Temp: 0.6) Entropy (Llama-3.2-3B-Instruct): Model: granite-4.1-8b-AWQ-INT4 (Temp: 1.25)

Classifier: Entropy (Llama-3.2-3B)

Performance:

Subset N AUROC TPR FPR Accuracy F1
Overall 41,784 0.4869 0.0005 0.0049 0.4978 0.0010
Prompt: direct_reference 10,356 0.4755 0.0004 0.0062 0.4971 0.0008
Prompt: indirect_reference 10,182 0.4783 0.0010 0.0037 0.4986 0.0020
Prompt: revise 10,666 0.5343 0.0000 0.0054 0.4973 0.0000
Prompt: rewrite 10,580 0.4567 0.0008 0.0042 0.4983 0.0015
✔️ Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 0.6) 6,064 0.6379 0.0003 0.0030 0.4987 0.0007
❗ Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 1.25) 6,004 0.2151 0.0003 0.0057 0.4973 0.0007
Model: Ministral-3-8B-Instruct-2512-AWQ-4bit (Temp: 0.6) 5,860 0.4974 0.0000 0.0061 0.4969 0.0000
Model: Qwen3-8B-AWQ (Temp: 0.7) 6,118 0.6106 0.0013 0.0042 0.4985 0.0026
Model: gemma-4-E4B-it (Temp: 0.7) 5,972 0.4135 0.0003 0.0027 0.4988 0.0007
Model: granite-4.1-8b-AWQ-INT4 (Temp: 0.6) 5,876 0.5525 0.0010 0.0068 0.4971 0.0020
Model: granite-4.1-8b-AWQ-INT4 (Temp: 1.25) 5,890 0.4814 0.0003 0.0058 0.4973 0.0007

Thresholding:

  • Direction: lower_is_ai
  • Swept for fpr_0_5pct with a found threshold of 0.2888.

Threshold Sweep: Entropy (Llama-3.2-3B)

Classification Histograms:

Classifier: Entropy (Llama-3.2-3B)

Per Prompt Subset

Entropy (Llama-3.2-3B): Prompt: direct_reference Entropy (Llama-3.2-3B): Prompt: indirect_reference Entropy (Llama-3.2-3B): Prompt: revise Entropy (Llama-3.2-3B): Prompt: rewrite

Per Generator Config Subset

Entropy (Llama-3.2-3B): Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 0.6) Entropy (Llama-3.2-3B): Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 1.25) Entropy (Llama-3.2-3B): Model: Ministral-3-8B-Instruct-2512-AWQ-4bit (Temp: 0.6) Entropy (Llama-3.2-3B): Model: Qwen3-8B-AWQ (Temp: 0.7) Entropy (Llama-3.2-3B): Model: gemma-4-E4B-it (Temp: 0.7) Entropy (Llama-3.2-3B): Model: granite-4.1-8b-AWQ-INT4 (Temp: 0.6) Entropy (Llama-3.2-3B): Model: granite-4.1-8b-AWQ-INT4 (Temp: 1.25)

Classifier: Top-p Outliers (Llama-3.2-3B-Instruct)

Performance:

Subset N AUROC TPR FPR Accuracy F1
Overall 41,784 0.5598 0.0108 0.0030 0.5039 0.0213
Prompt: direct_reference 10,356 0.5837 0.0160 0.0029 0.5066 0.0315
Prompt: indirect_reference 10,182 0.5977 0.0246 0.0029 0.5108 0.0478
Prompt: revise 10,666 0.5320 0.0017 0.0036 0.4991 0.0034
Prompt: rewrite 10,580 0.5280 0.0017 0.0026 0.4995 0.0034
✔️ Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 0.6) 6,064 0.6963 0.0501 0.0030 0.5236 0.0952
Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 1.25) 6,004 0.5704 0.0043 0.0040 0.5002 0.0086
Model: Ministral-3-8B-Instruct-2512-AWQ-4bit (Temp: 0.6) 5,860 0.5563 0.0010 0.0027 0.4991 0.0020
Model: Qwen3-8B-AWQ (Temp: 0.7) 6,118 0.6147 0.0111 0.0026 0.5042 0.0219
Model: gemma-4-E4B-it (Temp: 0.7) 5,972 0.5118 0.0044 0.0020 0.5012 0.0087
Model: granite-4.1-8b-AWQ-INT4 (Temp: 0.6) 5,876 0.5298 0.0027 0.0024 0.5002 0.0054
❗ Model: granite-4.1-8b-AWQ-INT4 (Temp: 1.25) 5,890 0.4325 0.0010 0.0044 0.4983 0.0020

Thresholding:

  • Direction: lower_is_ai
  • Swept for fpr_0_5pct with a found threshold of 0.0158.

Threshold Sweep: Top-p Outliers (Llama-3.2-3B-Instruct)

Classification Histograms:

Classifier: Top-p Outliers (Llama-3.2-3B-Instruct)

Per Prompt Subset

Top-p Outliers (Llama-3.2-3B-Instruct): Prompt: direct_reference Top-p Outliers (Llama-3.2-3B-Instruct): Prompt: indirect_reference Top-p Outliers (Llama-3.2-3B-Instruct): Prompt: revise Top-p Outliers (Llama-3.2-3B-Instruct): Prompt: rewrite

Per Generator Config Subset

Top-p Outliers (Llama-3.2-3B-Instruct): Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 0.6) Top-p Outliers (Llama-3.2-3B-Instruct): Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 1.25) Top-p Outliers (Llama-3.2-3B-Instruct): Model: Ministral-3-8B-Instruct-2512-AWQ-4bit (Temp: 0.6) Top-p Outliers (Llama-3.2-3B-Instruct): Model: Qwen3-8B-AWQ (Temp: 0.7) Top-p Outliers (Llama-3.2-3B-Instruct): Model: gemma-4-E4B-it (Temp: 0.7) Top-p Outliers (Llama-3.2-3B-Instruct): Model: granite-4.1-8b-AWQ-INT4 (Temp: 0.6) Top-p Outliers (Llama-3.2-3B-Instruct): Model: granite-4.1-8b-AWQ-INT4 (Temp: 1.25)

Classifier: Top-p Outliers (Llama-3.2-3B)

Performance:

Subset N AUROC TPR FPR Accuracy F1
Overall 41,784 0.5168 0.0062 0.0071 0.4995 0.0122
Prompt: direct_reference 10,356 0.6107 0.0100 0.0077 0.5012 0.0197
Prompt: indirect_reference 10,182 0.5249 0.0128 0.0063 0.5032 0.0251
Prompt: revise 10,666 0.4953 0.0009 0.0075 0.4967 0.0019
Prompt: rewrite 10,580 0.4396 0.0013 0.0068 0.4973 0.0026
✔️ Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 0.6) 6,064 0.6752 0.0168 0.0053 0.5058 0.0329
❗ Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 1.25) 6,004 0.2196 0.0010 0.0083 0.4963 0.0020
Model: Ministral-3-8B-Instruct-2512-AWQ-4bit (Temp: 0.6) 5,860 0.6009 0.0003 0.0075 0.4964 0.0007
Model: Qwen3-8B-AWQ (Temp: 0.7) 6,118 0.6662 0.0144 0.0072 0.5036 0.0282
Model: gemma-4-E4B-it (Temp: 0.7) 5,972 0.4851 0.0040 0.0040 0.5000 0.0080
Model: granite-4.1-8b-AWQ-INT4 (Temp: 0.6) 5,876 0.5588 0.0034 0.0085 0.4974 0.0067
Model: granite-4.1-8b-AWQ-INT4 (Temp: 1.25) 5,890 0.4100 0.0027 0.0088 0.4969 0.0054

Thresholding:

  • Direction: lower_is_ai
  • Swept for fpr_0_5pct with a found threshold of 0.0054.

Threshold Sweep: Top-p Outliers (Llama-3.2-3B)

Classification Histograms:

Classifier: Top-p Outliers (Llama-3.2-3B)

Per Prompt Subset

Top-p Outliers (Llama-3.2-3B): Prompt: direct_reference Top-p Outliers (Llama-3.2-3B): Prompt: indirect_reference Top-p Outliers (Llama-3.2-3B): Prompt: revise Top-p Outliers (Llama-3.2-3B): Prompt: rewrite

Per Generator Config Subset

Top-p Outliers (Llama-3.2-3B): Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 0.6) Top-p Outliers (Llama-3.2-3B): Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 1.25) Top-p Outliers (Llama-3.2-3B): Model: Ministral-3-8B-Instruct-2512-AWQ-4bit (Temp: 0.6) Top-p Outliers (Llama-3.2-3B): Model: Qwen3-8B-AWQ (Temp: 0.7) Top-p Outliers (Llama-3.2-3B): Model: gemma-4-E4B-it (Temp: 0.7) Top-p Outliers (Llama-3.2-3B): Model: granite-4.1-8b-AWQ-INT4 (Temp: 0.6) Top-p Outliers (Llama-3.2-3B): Model: granite-4.1-8b-AWQ-INT4 (Temp: 1.25)

Classifier: Top-k Outliers (Llama-3.2-3B-Instruct)

Performance:

Subset N AUROC TPR FPR Accuracy F1
Overall 41,784 0.5637 0.0191 0.0033 0.5079 0.0373
Prompt: direct_reference 10,356 0.6161 0.0373 0.0041 0.5166 0.0716
Prompt: indirect_reference 10,182 0.5662 0.0346 0.0018 0.5164 0.0667
Prompt: revise 10,666 0.5998 0.0038 0.0034 0.5002 0.0074
Prompt: rewrite 10,580 0.4725 0.0017 0.0038 0.4990 0.0034
✔️ Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 0.6) 6,064 0.7196 0.0627 0.0020 0.5303 0.1177
❗ Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 1.25) 6,004 0.2595 0.0013 0.0053 0.4980 0.0026
Model: Ministral-3-8B-Instruct-2512-AWQ-4bit (Temp: 0.6) 5,860 0.6421 0.0010 0.0048 0.4981 0.0020
Model: Qwen3-8B-AWQ (Temp: 0.7) 6,118 0.6911 0.0487 0.0023 0.5232 0.0927
Model: gemma-4-E4B-it (Temp: 0.7) 5,972 0.5050 0.0013 0.0013 0.5000 0.0027
Model: granite-4.1-8b-AWQ-INT4 (Temp: 0.6) 5,876 0.6051 0.0123 0.0044 0.5039 0.0241
Model: granite-4.1-8b-AWQ-INT4 (Temp: 1.25) 5,890 0.5226 0.0041 0.0027 0.5007 0.0081

Thresholding:

  • Direction: lower_is_ai
  • Swept for fpr_0_5pct with a found threshold of 0.0109.

Threshold Sweep: Top-k Outliers (Llama-3.2-3B-Instruct)

Classification Histograms:

Classifier: Top-k Outliers (Llama-3.2-3B-Instruct)

Per Prompt Subset

Top-k Outliers (Llama-3.2-3B-Instruct): Prompt: direct_reference Top-k Outliers (Llama-3.2-3B-Instruct): Prompt: indirect_reference Top-k Outliers (Llama-3.2-3B-Instruct): Prompt: revise Top-k Outliers (Llama-3.2-3B-Instruct): Prompt: rewrite

Per Generator Config Subset

Top-k Outliers (Llama-3.2-3B-Instruct): Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 0.6) Top-k Outliers (Llama-3.2-3B-Instruct): Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 1.25) Top-k Outliers (Llama-3.2-3B-Instruct): Model: Ministral-3-8B-Instruct-2512-AWQ-4bit (Temp: 0.6) Top-k Outliers (Llama-3.2-3B-Instruct): Model: Qwen3-8B-AWQ (Temp: 0.7) Top-k Outliers (Llama-3.2-3B-Instruct): Model: gemma-4-E4B-it (Temp: 0.7) Top-k Outliers (Llama-3.2-3B-Instruct): Model: granite-4.1-8b-AWQ-INT4 (Temp: 0.6) Top-k Outliers (Llama-3.2-3B-Instruct): Model: granite-4.1-8b-AWQ-INT4 (Temp: 1.25)

Classifier: Top-k Outliers (Llama-3.2-3B)

Performance:

Subset N AUROC TPR FPR Accuracy F1
Overall 41,784 0.5168 0.0053 0.0048 0.5002 0.0104
Prompt: direct_reference 10,356 0.5622 0.0085 0.0054 0.5015 0.0168
Prompt: indirect_reference 10,182 0.5113 0.0104 0.0039 0.5032 0.0205
Prompt: revise 10,666 0.5465 0.0011 0.0054 0.4978 0.0022
Prompt: rewrite 10,580 0.4465 0.0013 0.0043 0.4985 0.0026
✔️ Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 0.6) 6,064 0.6778 0.0155 0.0033 0.5061 0.0304
❗ Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 1.25) 6,004 0.2099 0.0007 0.0060 0.4973 0.0013
Model: Ministral-3-8B-Instruct-2512-AWQ-4bit (Temp: 0.6) 5,860 0.5863 0.0000 0.0055 0.4973 0.0000
Model: Qwen3-8B-AWQ (Temp: 0.7) 6,118 0.6659 0.0147 0.0039 0.5054 0.0289
Model: gemma-4-E4B-it (Temp: 0.7) 5,972 0.4514 0.0007 0.0027 0.4990 0.0013
Model: granite-4.1-8b-AWQ-INT4 (Temp: 0.6) 5,876 0.5570 0.0041 0.0071 0.4985 0.0081
Model: granite-4.1-8b-AWQ-INT4 (Temp: 1.25) 5,890 0.4683 0.0007 0.0051 0.4978 0.0014

Thresholding:

  • Direction: lower_is_ai
  • Swept for fpr_0_5pct with a found threshold of 0.0056.

Threshold Sweep: Top-k Outliers (Llama-3.2-3B)

Classification Histograms:

Classifier: Top-k Outliers (Llama-3.2-3B)

Per Prompt Subset

Top-k Outliers (Llama-3.2-3B): Prompt: direct_reference Top-k Outliers (Llama-3.2-3B): Prompt: indirect_reference Top-k Outliers (Llama-3.2-3B): Prompt: revise Top-k Outliers (Llama-3.2-3B): Prompt: rewrite

Per Generator Config Subset

Top-k Outliers (Llama-3.2-3B): Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 0.6) Top-k Outliers (Llama-3.2-3B): Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 1.25) Top-k Outliers (Llama-3.2-3B): Model: Ministral-3-8B-Instruct-2512-AWQ-4bit (Temp: 0.6) Top-k Outliers (Llama-3.2-3B): Model: Qwen3-8B-AWQ (Temp: 0.7) Top-k Outliers (Llama-3.2-3B): Model: gemma-4-E4B-it (Temp: 0.7) Top-k Outliers (Llama-3.2-3B): Model: granite-4.1-8b-AWQ-INT4 (Temp: 0.6) Top-k Outliers (Llama-3.2-3B): Model: granite-4.1-8b-AWQ-INT4 (Temp: 1.25)

Classifier: FastDetectGPT (Llama-3.2-3B-Instruct)

Performance:

Subset N AUROC TPR FPR Accuracy F1
Overall 41,784 0.5725 0.0027 0.0041 0.4993 0.0054
Prompt: direct_reference 10,356 0.5900 0.0023 0.0048 0.4987 0.0046
Prompt: indirect_reference 10,182 0.6299 0.0020 0.0045 0.4987 0.0039
Prompt: revise 10,666 0.5402 0.0034 0.0047 0.4993 0.0067
Prompt: rewrite 10,580 0.5334 0.0032 0.0025 0.5004 0.0064
✔️ Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 0.6) 6,064 0.6945 0.0013 0.0053 0.4980 0.0026
Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 1.25) 6,004 0.5419 0.0080 0.0067 0.5007 0.0158
Model: Ministral-3-8B-Instruct-2512-AWQ-4bit (Temp: 0.6) 5,860 0.5923 0.0007 0.0027 0.4990 0.0014
Model: Qwen3-8B-AWQ (Temp: 0.7) 6,118 0.5891 0.0010 0.0016 0.4997 0.0020
Model: gemma-4-E4B-it (Temp: 0.7) 5,972 0.5690 0.0044 0.0057 0.4993 0.0086
Model: granite-4.1-8b-AWQ-INT4 (Temp: 0.6) 5,876 0.5443 0.0017 0.0044 0.4986 0.0034
❗ Model: granite-4.1-8b-AWQ-INT4 (Temp: 1.25) 5,890 0.4715 0.0020 0.0024 0.4998 0.0041

Thresholding:

  • Direction: higher_is_ai
  • Swept for fpr_0_5pct with a found threshold of 6.7770.

Threshold Sweep: FastDetectGPT (Llama-3.2-3B-Instruct)

Classification Histograms:

Classifier: FastDetectGPT (Llama-3.2-3B-Instruct)

Per Prompt Subset

FastDetectGPT (Llama-3.2-3B-Instruct): Prompt: direct_reference FastDetectGPT (Llama-3.2-3B-Instruct): Prompt: indirect_reference FastDetectGPT (Llama-3.2-3B-Instruct): Prompt: revise FastDetectGPT (Llama-3.2-3B-Instruct): Prompt: rewrite

Per Generator Config Subset

FastDetectGPT (Llama-3.2-3B-Instruct): Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 0.6) FastDetectGPT (Llama-3.2-3B-Instruct): Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 1.25) FastDetectGPT (Llama-3.2-3B-Instruct): Model: Ministral-3-8B-Instruct-2512-AWQ-4bit (Temp: 0.6) FastDetectGPT (Llama-3.2-3B-Instruct): Model: Qwen3-8B-AWQ (Temp: 0.7) FastDetectGPT (Llama-3.2-3B-Instruct): Model: gemma-4-E4B-it (Temp: 0.7) FastDetectGPT (Llama-3.2-3B-Instruct): Model: granite-4.1-8b-AWQ-INT4 (Temp: 0.6) FastDetectGPT (Llama-3.2-3B-Instruct): Model: granite-4.1-8b-AWQ-INT4 (Temp: 1.25)

Classifier: FastDetectGPT (Llama-3.2-3B)

Performance:

Subset N AUROC TPR FPR Accuracy F1
Overall 41,784 0.5159 0.0986 0.0066 0.5460 0.1784
Prompt: direct_reference 10,356 0.4288 0.0709 0.0070 0.5320 0.1315
Prompt: indirect_reference 10,182 0.4831 0.1159 0.0075 0.5542 0.2063
Prompt: revise 10,666 0.5541 0.1001 0.0056 0.5473 0.1811
Prompt: rewrite 10,580 0.5944 0.1076 0.0062 0.5507 0.1931
❗ Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 0.6) 6,064 0.2939 0.0030 0.0082 0.4974 0.0059
✔️ Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 1.25) 6,004 0.8992 0.5673 0.0057 0.7808 0.7213
Model: Ministral-3-8B-Instruct-2512-AWQ-4bit (Temp: 0.6) 5,860 0.3979 0.0143 0.0075 0.5034 0.0281
Model: Qwen3-8B-AWQ (Temp: 0.7) 6,118 0.3076 0.0082 0.0062 0.5010 0.0161
Model: gemma-4-E4B-it (Temp: 0.7) 5,972 0.5519 0.0131 0.0064 0.5033 0.0256
Model: granite-4.1-8b-AWQ-INT4 (Temp: 0.6) 5,876 0.4675 0.0099 0.0054 0.5022 0.0194
Model: granite-4.1-8b-AWQ-INT4 (Temp: 1.25) 5,890 0.6924 0.0723 0.0065 0.5329 0.1341

Thresholding:

  • Direction: lower_is_ai
  • Swept for fpr_0_5pct with a found threshold of -3.2326.

Threshold Sweep: FastDetectGPT (Llama-3.2-3B)

Classification Histograms:

Classifier: FastDetectGPT (Llama-3.2-3B)

Per Prompt Subset

FastDetectGPT (Llama-3.2-3B): Prompt: direct_reference FastDetectGPT (Llama-3.2-3B): Prompt: indirect_reference FastDetectGPT (Llama-3.2-3B): Prompt: revise FastDetectGPT (Llama-3.2-3B): Prompt: rewrite

Per Generator Config Subset

FastDetectGPT (Llama-3.2-3B): Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 0.6) FastDetectGPT (Llama-3.2-3B): Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 1.25) FastDetectGPT (Llama-3.2-3B): Model: Ministral-3-8B-Instruct-2512-AWQ-4bit (Temp: 0.6) FastDetectGPT (Llama-3.2-3B): Model: Qwen3-8B-AWQ (Temp: 0.7) FastDetectGPT (Llama-3.2-3B): Model: gemma-4-E4B-it (Temp: 0.7) FastDetectGPT (Llama-3.2-3B): Model: granite-4.1-8b-AWQ-INT4 (Temp: 0.6) FastDetectGPT (Llama-3.2-3B): Model: granite-4.1-8b-AWQ-INT4 (Temp: 1.25)

Classifier: Binoculars

Performance:

Subset N AUROC TPR FPR Accuracy F1
Overall 41,784 0.6302 0.0121 0.0031 0.5045 0.0239
Prompt: direct_reference 10,356 0.6006 0.0114 0.0027 0.5043 0.0225
Prompt: indirect_reference 10,182 0.6500 0.0192 0.0037 0.5078 0.0376
Prompt: revise 10,666 0.6654 0.0075 0.0034 0.5021 0.0148
Prompt: rewrite 10,580 0.6046 0.0106 0.0026 0.5040 0.0209
Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 0.6) 6,064 0.5311 0.0013 0.0036 0.4988 0.0026
✔️ Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 1.25) 6,004 0.8816 0.0403 0.0033 0.5185 0.0772
Model: Ministral-3-8B-Instruct-2512-AWQ-4bit (Temp: 0.6) 5,860 0.5429 0.0092 0.0038 0.5027 0.0182
❗ Model: Qwen3-8B-AWQ (Temp: 0.7) 6,118 0.4838 0.0013 0.0036 0.4989 0.0026
Model: gemma-4-E4B-it (Temp: 0.7) 5,972 0.6467 0.0074 0.0033 0.5020 0.0146
Model: granite-4.1-8b-AWQ-INT4 (Temp: 0.6) 5,876 0.6074 0.0065 0.0017 0.5024 0.0128
Model: granite-4.1-8b-AWQ-INT4 (Temp: 1.25) 5,890 0.7140 0.0190 0.0024 0.5083 0.0372

Thresholding:

  • Direction: higher_is_ai
  • Swept for fpr_0_5pct with a found threshold of 1.0472.

Threshold Sweep: Binoculars

Classification Histograms:

Classifier: Binoculars

Per Prompt Subset

Binoculars: Prompt: direct_reference Binoculars: Prompt: indirect_reference Binoculars: Prompt: revise Binoculars: Prompt: rewrite

Per Generator Config Subset

Binoculars: Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 0.6) Binoculars: Model: Meta-Llama-3.1-8B-Instruct-AWQ-INT4 (Temp: 1.25) Binoculars: Model: Ministral-3-8B-Instruct-2512-AWQ-4bit (Temp: 0.6) Binoculars: Model: Qwen3-8B-AWQ (Temp: 0.7) Binoculars: Model: gemma-4-E4B-it (Temp: 0.7) Binoculars: Model: granite-4.1-8b-AWQ-INT4 (Temp: 0.6) Binoculars: Model: granite-4.1-8b-AWQ-INT4 (Temp: 1.25)

Downloads last month
507