Add annotations with Gemini-2.5-Flash

.gitattributes

@@ -57,6 +57,7 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 # Video files - compressed
 *.mp4 filter=lfs diff=lfs merge=lfs -text
 *.webm filter=lfs diff=lfs merge=lfs -text
+*.json filter=lfs diff=lfs merge=lfs -text
 allenai_multipref.json filter=lfs diff=lfs merge=lfs -text
 chatbot_arena.json filter=lfs diff=lfs merge=lfs -text
 model_comparison.json filter=lfs diff=lfs merge=lfs -text

README.md

@@ -1,6 +1,6 @@
 ---
 license: odc-by
-pretty_name: Feedback Forensics Annotations v1.0
+pretty_name: Feedback Forensics Annotations v2.0
 configs:
 - config_name: default
   data_files:
@@ -18,23 +18,23 @@ configs:
 
 # Feedback Forensics Annotations
 
-This dataset contains the personality annotations from the experiments included in the *Feedback Forensics* paper. 
-The annotations are provided over pairwise model outputs, typically consisting of a prompt and two responses. 
-Each individual annotation considers as single personality trait (e.g. "confidence"). The annotation indicates when 
-two responses differ with respect to that trait 
-(e.g. one response is more confident), a trait does not apply to both responses or is equal in both responses. 
-Annotations were collected using the [Inverse Constitutional AI (ICAI)](https://github.com/rdnfn/icai) annotation pipeline with a 
-fixed set of personality traits to be tested, using `gpt-4o-mini-2024-07-18`. The data is provided in the AnnotatedPairs json data 
-format, [described here](https://github.com/rdnfn/icai?tab=readme-ov-file#using-results-for-further-analysis-the-annotated-pairs-json-format-v20). 
+This dataset contains the personality annotations from the experiments included in the *Feedback Forensics* paper.
+The annotations are provided over pairwise model outputs, typically consisting of a prompt and two responses.
+Each individual annotation considers as single personality trait (e.g. "confidence"). The annotation indicates when
+two responses differ with respect to that trait
+(e.g. one response is more confident), a trait does not apply to both responses or is equal in both responses.
+Annotations were collected using the [Inverse Constitutional AI (ICAI)](https://github.com/rdnfn/icai) annotation pipeline with a
+fixed set of personality traits to be tested, using `gemini-2.5-flash`. The data is provided in the AnnotatedPairs json data
+format, [described here](https://github.com/rdnfn/icai?tab=readme-ov-file#using-results-for-further-analysis-the-annotated-pairs-json-format-v20).
 
-Note that we do *not* include prompts and responses from the original datasets, instead providing metadata 
-(e.g., `conversation_id`) to enable merging with the base data. The annotation data is 
+Note that we do *not* include prompts and responses from the original datasets, instead providing metadata
+(e.g., `conversation_id`) to enable merging with the base data. The annotation data is
 sufficient for independent local analysis with the Feedback Forensics Gradio app, even without merging.
 
 ## Format
 
-The data is provided in the AnnotatedPairs json data 
-format, [described here](https://github.com/rdnfn/icai?tab=readme-ov-file#using-results-for-further-analysis-the-annotated-pairs-json-format-v20). 
+The data is provided in the AnnotatedPairs json data
+format, [described here](https://github.com/rdnfn/icai?tab=readme-ov-file#using-results-for-further-analysis-the-annotated-pairs-json-format-v20).
 See a minimal illustrative example below:
 
 ```json
@@ -92,8 +92,8 @@ See a minimal illustrative example below:
 }
 ```
 
-The main annotations are included under `dataset['comparisons'][i]["annotations"]`, 
-where i is the index of the annotation. `dataset['comparisons'][i]["metadata"]` 
+The main annotations are included under `dataset['comparisons'][i]["annotations"]`,
+where i is the index of the annotation. `dataset['comparisons'][i]["metadata"]`
 provides metadata to enable merging.
 
 ## Annotated datsets

data/allenai_multipref.json

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:07ddfd92cd3e22cea41741dd472bb7b32656e70ac285d9a065e7402035ac173b
-size 61560143
+oid sha256:7be407c3061d8c4dc636ff5ec5d4816aed198da0a5f06455b8415b24642cd5d4
+size 64149358

data/archive/gpt4omini_annotation_data/allenai_multipref.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:07ddfd92cd3e22cea41741dd472bb7b32656e70ac285d9a065e7402035ac173b
+size 61560143

data/archive/gpt4omini_annotation_data/chatbot_arena.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d424f6143af912da11e6798ae32ad92d833afde7d131623fae2b8b0b43da63fd
+size 56727399

data/archive/gpt4omini_annotation_data/llama4_arena_vs_public_version.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e081379675ae2c25922d99abefd986567ad03381f63842970abfc45178a8bb6d
+size 34352216

data/archive/gpt4omini_annotation_data/model_comparison.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:db8afce12e15453ab8ab8480e5357b159df6b4d02b80fc8bc58ce24e5f489de4
+size 24914518

data/archive/gpt4omini_annotation_data/prism.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1d8f8b6a4369f1b7b6c0c70c4f0ca7f230f909a4e93f869d19fc1921d258de7d
+size 43987218

data/chatbot_arena.json

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d424f6143af912da11e6798ae32ad92d833afde7d131623fae2b8b0b43da63fd
-size 56727399
+oid sha256:5bfafbec79ec6f868312ab446f0ad612626832c6ee5421d23a6689ab4b675f30
+size 58150304

data/llama4_arena_vs_public_version.json

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e081379675ae2c25922d99abefd986567ad03381f63842970abfc45178a8bb6d
-size 34352216
+oid sha256:c1dbd47fb2c9e715fde9c4db8e7f582596d8aead1cdd98b243589f04fea0f4a6
+size 16384302

data/model_comparison.json

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:db8afce12e15453ab8ab8480e5357b159df6b4d02b80fc8bc58ce24e5f489de4
-size 24914518
+oid sha256:cb1e374b8cc29ca9b551da5bd409a1f20011ebfe944221f4977fba96b0600c10
+size 28569963

data/prism.json

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1d8f8b6a4369f1b7b6c0c70c4f0ca7f230f909a4e93f869d19fc1921d258de7d
-size 43987218
+oid sha256:57198d3ec2c92fe352631483517116fd2de299d4c7bfe8a46a507606e45f8718
+size 45392464