condafile commited on
Commit
d41769e
·
verified ·
1 Parent(s): 486063c

Add Figure 1 and third-party licensing notice

Browse files
Files changed (4) hide show
  1. .gitattributes +1 -0
  2. README.md +9 -0
  3. figure1.pdf +3 -0
  4. figure1.png +3 -0
.gitattributes CHANGED
@@ -58,3 +58,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
58
  # Video files - compressed
59
  *.mp4 filter=lfs diff=lfs merge=lfs -text
60
  *.webm filter=lfs diff=lfs merge=lfs -text
 
 
58
  # Video files - compressed
59
  *.mp4 filter=lfs diff=lfs merge=lfs -text
60
  *.webm filter=lfs diff=lfs merge=lfs -text
61
+ figure1.pdf filter=lfs diff=lfs merge=lfs -text
README.md CHANGED
@@ -43,6 +43,11 @@ configs:
43
 
44
  The evaluation dataset for **When Protection Overrides Consequences: Understanding Protection Mode in Large Language Model Decision-Making**. Main compares a low-immediate-risk person/animal with an object facing larger time-sensitive physical damage. Reference labels describe the larger stated physical consequence, not universal moral correctness.
45
 
 
 
 
 
 
46
  ## Load immediately
47
 
48
  ```python
@@ -135,3 +140,7 @@ Main metrics are **GCR** (object-choice response rate), **OAR** (subject-choice
135
  Neutral exposes its vulnerable-object description rather than assigning a living-subject label. It should be judged by the two objects named in that prompt.
136
 
137
  `DATA_MANIFEST.json` records each Parquet file's row count, byte size, and SHA-256. Conversion preserves every prompt byte and identifier from the frozen release. These are test/evaluation sets, with no training split. Dataset content is CC BY 4.0; the standalone evaluation code is MIT-licensed in the linked code repository.
 
 
 
 
 
43
 
44
  The evaluation dataset for **When Protection Overrides Consequences: Understanding Protection Mode in Large Language Model Decision-Making**. Main compares a low-immediate-risk person/animal with an object facing larger time-sensitive physical damage. Reference labels describe the larger stated physical consequence, not universal moral correctness.
45
 
46
+
47
+ ![Figure 1: Protection Mode examples and model-level results](figure1.png)
48
+
49
+ *Figure 1. Protection Mode examples and results for subject-priority coverage and knowledge–action divergence.* [Download the original PDF](figure1.pdf).
50
+
51
  ## Load immediately
52
 
53
  ```python
 
140
  Neutral exposes its vulnerable-object description rather than assigning a living-subject label. It should be judged by the two objects named in that prompt.
141
 
142
  `DATA_MANIFEST.json` records each Parquet file's row count, byte size, and SHA-256. Conversion preserves every prompt byte and identifier from the frozen release. These are test/evaluation sets, with no training split. Dataset content is CC BY 4.0; the standalone evaluation code is MIT-licensed in the linked code repository.
143
+
144
+ ## Third-party models and services
145
+
146
+ Unless explicitly stated otherwise, this repository does not include or distribute any third-party model weights. Users who independently download or enable third-party models or services must also comply with the applicable upstream project, model-weight, and data license terms. Use of third-party API services is subject to the applicable service providers' terms and conditions.
figure1.pdf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9aff28f378a6c3cc1453a09283d55bea48e964e8a559afdc889c57f4b785654d
3
+ size 516996
figure1.png ADDED

Git LFS Details

  • SHA256: 09120d764084ac62ce690056153e3f2bae4d006f7af8eebc0668c16fcd714ac8
  • Pointer size: 132 Bytes
  • Size of remote file: 2.59 MB