Commit ·
76fadcc
0
Parent(s):
Duplicate from TylerZ0931/PACIFIC-big-five-trait-preferences
Browse files- .gitattributes +61 -0
- PACIFIC.jsonl +3 -0
- README.md +139 -0
.gitattributes
ADDED
|
@@ -0,0 +1,61 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
*.7z filter=lfs diff=lfs merge=lfs -text
|
| 2 |
+
*.arrow filter=lfs diff=lfs merge=lfs -text
|
| 3 |
+
*.avro filter=lfs diff=lfs merge=lfs -text
|
| 4 |
+
*.bin filter=lfs diff=lfs merge=lfs -text
|
| 5 |
+
*.bz2 filter=lfs diff=lfs merge=lfs -text
|
| 6 |
+
*.ckpt filter=lfs diff=lfs merge=lfs -text
|
| 7 |
+
*.ftz filter=lfs diff=lfs merge=lfs -text
|
| 8 |
+
*.gz filter=lfs diff=lfs merge=lfs -text
|
| 9 |
+
*.h5 filter=lfs diff=lfs merge=lfs -text
|
| 10 |
+
*.joblib filter=lfs diff=lfs merge=lfs -text
|
| 11 |
+
*.lfs.* filter=lfs diff=lfs merge=lfs -text
|
| 12 |
+
*.lz4 filter=lfs diff=lfs merge=lfs -text
|
| 13 |
+
*.mds filter=lfs diff=lfs merge=lfs -text
|
| 14 |
+
*.mlmodel filter=lfs diff=lfs merge=lfs -text
|
| 15 |
+
*.model filter=lfs diff=lfs merge=lfs -text
|
| 16 |
+
*.msgpack filter=lfs diff=lfs merge=lfs -text
|
| 17 |
+
*.npy filter=lfs diff=lfs merge=lfs -text
|
| 18 |
+
*.npz filter=lfs diff=lfs merge=lfs -text
|
| 19 |
+
*.onnx filter=lfs diff=lfs merge=lfs -text
|
| 20 |
+
*.ot filter=lfs diff=lfs merge=lfs -text
|
| 21 |
+
*.parquet filter=lfs diff=lfs merge=lfs -text
|
| 22 |
+
*.pb filter=lfs diff=lfs merge=lfs -text
|
| 23 |
+
*.pickle filter=lfs diff=lfs merge=lfs -text
|
| 24 |
+
*.pkl filter=lfs diff=lfs merge=lfs -text
|
| 25 |
+
*.pt filter=lfs diff=lfs merge=lfs -text
|
| 26 |
+
*.pth filter=lfs diff=lfs merge=lfs -text
|
| 27 |
+
*.rar filter=lfs diff=lfs merge=lfs -text
|
| 28 |
+
*.safetensors filter=lfs diff=lfs merge=lfs -text
|
| 29 |
+
saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
| 30 |
+
*.tar.* filter=lfs diff=lfs merge=lfs -text
|
| 31 |
+
*.tar filter=lfs diff=lfs merge=lfs -text
|
| 32 |
+
*.tflite filter=lfs diff=lfs merge=lfs -text
|
| 33 |
+
*.tgz filter=lfs diff=lfs merge=lfs -text
|
| 34 |
+
*.wasm filter=lfs diff=lfs merge=lfs -text
|
| 35 |
+
*.xz filter=lfs diff=lfs merge=lfs -text
|
| 36 |
+
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 37 |
+
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 38 |
+
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
| 39 |
+
# Audio files - uncompressed
|
| 40 |
+
*.pcm filter=lfs diff=lfs merge=lfs -text
|
| 41 |
+
*.sam filter=lfs diff=lfs merge=lfs -text
|
| 42 |
+
*.raw filter=lfs diff=lfs merge=lfs -text
|
| 43 |
+
# Audio files - compressed
|
| 44 |
+
*.aac filter=lfs diff=lfs merge=lfs -text
|
| 45 |
+
*.flac filter=lfs diff=lfs merge=lfs -text
|
| 46 |
+
*.mp3 filter=lfs diff=lfs merge=lfs -text
|
| 47 |
+
*.ogg filter=lfs diff=lfs merge=lfs -text
|
| 48 |
+
*.wav filter=lfs diff=lfs merge=lfs -text
|
| 49 |
+
# Image files - uncompressed
|
| 50 |
+
*.bmp filter=lfs diff=lfs merge=lfs -text
|
| 51 |
+
*.gif filter=lfs diff=lfs merge=lfs -text
|
| 52 |
+
*.png filter=lfs diff=lfs merge=lfs -text
|
| 53 |
+
*.tiff filter=lfs diff=lfs merge=lfs -text
|
| 54 |
+
# Image files - compressed
|
| 55 |
+
*.jpg filter=lfs diff=lfs merge=lfs -text
|
| 56 |
+
*.jpeg filter=lfs diff=lfs merge=lfs -text
|
| 57 |
+
*.webp 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 |
+
PACIFIC.jsonl filter=lfs diff=lfs merge=lfs -text
|
PACIFIC.jsonl
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8cb201b7a6b894db635b229eafe94e803497ce09494553cf7b295be68f2128e0
|
| 3 |
+
size 17634753
|
README.md
ADDED
|
@@ -0,0 +1,139 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
language: en
|
| 4 |
+
tags:
|
| 5 |
+
- personality
|
| 6 |
+
- preference
|
| 7 |
+
- personalization
|
| 8 |
+
- memory
|
| 9 |
+
---
|
| 10 |
+
Dataset For Paper "Can LLMs Discern the Traits Influencing Your Preferences? Evaluating Personality-Driven Preference Alignment in LLMs"
|
| 11 |
+
|
| 12 |
+
**PACIFIC** (*Preference Alignment for Choices Inference via Five-factor Identity Characterization*) is a psychometrics-grounded dataset for studying whether Large Language Models can use **stable personality traits** — rather than exhaustive preference logs — as a latent signal for inferring user preferences on unseen queries.
|
| 13 |
+
|
| 14 |
+
It contains **1,200 preference–query pairs** across **20 everyday domains**, each with four candidate answers and full **Big-Five (OCEAN)** annotations on both the stated preference and every answer choice. The dataset is balanced so that all **ten high/low trait directions** are evenly represented (120 pairs each),
|
| 15 |
+
|
| 16 |
+
## Why this dataset
|
| 17 |
+
|
| 18 |
+
As user history grows, models get *worse* at following stated preferences — prior work reports preference-following accuracy falling below 10% after ~10 turns. Humans do the opposite: they abstract many small observations into a stable sense of a person and generalize from it. PACIFIC operationalizes that idea by asking whether Big-Five traits can serve as a compact latent representation that connects preferences across domains through shared psychometric alignment rather than surface-level topical similarity.
|
| 19 |
+
|
| 20 |
+
Each item is deliberately constructed as a **hard case**: with the preference removed, a strong LLM scores at roughly chance (25%), because all four options look equally reasonable from the query alone. The correct answer only becomes clear once the user's personality-driven preference is known.
|
| 21 |
+
|
| 22 |
+
## Dataset structure
|
| 23 |
+
|
| 24 |
+
Single split, JSON format:
|
| 25 |
+
|
| 26 |
+
| | |
|
| 27 |
+
|---|---|
|
| 28 |
+
| Split | `train` |
|
| 29 |
+
| Rows | 1,200 |
|
| 30 |
+
| Trait directions | 10 (High/Low × O, C, E, A, N), 120 rows each |
|
| 31 |
+
| Domains | 20 |
|
| 32 |
+
| Size | ~17.6 MB |
|
| 33 |
+
| Language | English |
|
| 34 |
+
|
| 35 |
+
### Fields
|
| 36 |
+
|
| 37 |
+
| Field | Type | Description |
|
| 38 |
+
|---|---|---|
|
| 39 |
+
| `trait` | string | Target Big-Five direction for the item, e.g. `High Openness`, `Low Neuroticism` (one of the ten). |
|
| 40 |
+
| `topic` | string | Domain the scenario belongs to, e.g. `Home_Cooking`, `Personal_Finance` (one of 20). |
|
| 41 |
+
| `preference` | string | The user's personality-driven preference statement. |
|
| 42 |
+
| `question` | string | The preference-sensitive user query. Answerable in general, but a generic answer tends to violate the preference. |
|
| 43 |
+
| `explanation` | string | Difficulty rationale: why a default/"safe" response would fail to respect the preference. |
|
| 44 |
+
| `options` | list[string] | The four candidate response texts (plain strings). |
|
| 45 |
+
| `preference_statement` | dict | Full OCEAN annotation of the preference (see below). |
|
| 46 |
+
| `choices` | list[dict] | The four candidate answers with per-choice OCEAN annotations. |
|
| 47 |
+
| `ground_truth_choice_id` | int | Index of the preference-adhering choice. |
|
| 48 |
+
|
| 49 |
+
### Nested annotation schema
|
| 50 |
+
|
| 51 |
+
Both `preference_statement` and each element of `choices` carry Big-Five annotations produced during dataset construction:
|
| 52 |
+
|
| 53 |
+
```jsonc
|
| 54 |
+
// preference_statement
|
| 55 |
+
{
|
| 56 |
+
"text": "I find following traditional recipes step-by-step incredibly stifling ...",
|
| 57 |
+
"related_to_personality": 1,
|
| 58 |
+
"openness": 7,
|
| 59 |
+
"openness_explanation": "The user's preference for improvising with 'avant-garde' combinations ...",
|
| 60 |
+
// ... conscientiousness, extraversion, agreeableness, neuroticism,
|
| 61 |
+
// each with an integer score (1–7 Likert) and a matching *_explanation
|
| 62 |
+
}
|
| 63 |
+
```
|
| 64 |
+
|
| 65 |
+
```jsonc
|
| 66 |
+
// one entry in choices[]
|
| 67 |
+
{
|
| 68 |
+
"choice_id": 0,
|
| 69 |
+
"text": "Use a basic apple pie framework but innovate by infusing ...",
|
| 70 |
+
"related_to_personality": 1,
|
| 71 |
+
"openness": 7,
|
| 72 |
+
"openness_explanation": "The user explicitly wants to innovate ...",
|
| 73 |
+
// ... remaining OCEAN traits with scores + explanations
|
| 74 |
+
}
|
| 75 |
+
```
|
| 76 |
+
|
| 77 |
+
Trait scores use a **7-point Likert scale** (1 = very low expression, 7 = very high). For downstream label mapping the paper discretizes these as *low* (<4), *high* (>4), or *unclear* (=4). See the paper's supplement for the exact annotation rubric, including per-trait confidence weighting.
|
| 78 |
+
|
| 79 |
+
### ⚠️ Evaluation: option ordering
|
| 80 |
+
|
| 81 |
+
In the released items the **preference-adhering option is generated first** (`ground_truth_choice_id` is `0`, i.e. `choices[0]`), with the three distractors following. When evaluating a model, **shuffle the option order per item** (and re-map the ground truth) to avoid rewarding position bias. Also note the dataset stores a **0-based** ground-truth index; if you follow the paper's prompt format, which asks the model to return an integer in `{1,2,3,4}`, convert accordingly.
|
| 82 |
+
|
| 83 |
+
|
| 84 |
+
## Intended uses
|
| 85 |
+
|
| 86 |
+
- **Preference-following / personalization evaluation** — can a model apply a latent preference constraint instead of defaulting to generic helpfulness?
|
| 87 |
+
- **Cross-domain preference inference** — testing whether trait-aligned preferences from *other* domains help answer a query whose preference was never explicitly stated.
|
| 88 |
+
- **Personality-informed retrieval** — building and evaluating retrievers that surface trait-consistent context (e.g. the paper's persona-aware contrastive retriever, PiRAG).
|
| 89 |
+
- **Trait prediction from behavior** — predicting OCEAN traits from preferences or from chosen answers.
|
| 90 |
+
|
| 91 |
+
## Dataset creation
|
| 92 |
+
|
| 93 |
+
Items were synthesized with Gemini 2.5 Pro over the 20-topic taxonomy shared with PrefEval, but **anchored to psychometric traits** so coverage is balanced across all ten high/low directions rather than derived directly from topics.
|
| 94 |
+
|
| 95 |
+
**Generation.** For each item the pipeline produces a personality-driven preference, a preference-sensitive query, a difficulty rationale, and four candidate responses (one adhering, three plausible-but-misaligned distractors). A **High-Violation constraint** forces queries where a generic answer is likely to break the stated preference, so surface keyword matching is insufficient.
|
| 96 |
+
|
| 97 |
+
**Annotation.** Every preference and every choice is scored on all five OCEAN dimensions using a **Dual-Strategy Annotation Protocol**:
|
| 98 |
+
- **Mirroring strategy (O, C, E, A)** — grounded in Self-Congruity Theory; high-trait users are annotated toward preferences that reinforce that trait (e.g. a highly Open user favors novelty/complexity).
|
| 99 |
+
- **Compensatory strategy (N)** — Neuroticism is annotated along a *Safety vs. Efficiency* trade-off, since high-N users seek reassurance/predictability rather than "more neurotic" experiences, while low-N users prioritize efficiency.
|
| 100 |
+
|
| 101 |
+
**Quality control.** An independent judge model (GPT-4o-mini) retained only items where the ground truth optimally aligns with the preference while all distractors violate or ignore it, and where the preference logically reflects its assigned trait. A confidence threshold (τ = 0.7) filters uncertain pairs.
|
| 102 |
+
|
| 103 |
+
**Human validation.** 15 annotators from varied professional backgrounds completed a Big-Five test, then validated items matched to their own dominant trait. On trait-matched choice selection they reached ~78% average accuracy with high inter-annotator agreement (Fleiss' κ ≈ 0.86), and reported a 91% approval rate that the ground-truth option was the best fit. The automated evaluator agreed strongly with the human consensus (Cohen's κ ≈ 0.92).
|
| 104 |
+
|
| 105 |
+
## Baseline findings (from the paper)
|
| 106 |
+
|
| 107 |
+
- With the preference **withheld**, models sit near chance (~25%), confirming the distractors are genuinely hard.
|
| 108 |
+
- **Trait-aligned** preference context substantially outperforms mixed-trait context, and adding unrelated preferences (contamination) reduces accuracy — organizing history by shared traits helps; raw volume does not (expanding mixed context from 5→25 preferences slightly *hurt* accuracy).
|
| 109 |
+
- Given clean trait-aligned context, strong models approach ceiling (Gemini-2.5-Pro up to ~99%), indicating the bottleneck is **context construction / retrieval**, not reasoning.
|
| 110 |
+
- Without trait labels, a plain dense retriever reaches ~30% while persona-aware contrastive fine-tuning (**PiRAG**) raises label-free accuracy to ~43%.
|
| 111 |
+
- A **social-desirability bias** appears when predicting traits *about users*: models identify "high" traits well but collapse on "low" directions, likely an RLHF artifact.
|
| 112 |
+
|
| 113 |
+
## Limitations
|
| 114 |
+
|
| 115 |
+
- Traits are simplified to **binary High/Low** rather than continuous Likert values.
|
| 116 |
+
- Evaluation is **multiple-choice QA**, not open-ended generation.
|
| 117 |
+
- Preferences and manifestations are treated as **culture- and language-independent** (English only).
|
| 118 |
+
- Content is **LLM-synthesized** (with human validation), so residual generator artifacts may remain.
|
| 119 |
+
|
| 120 |
+
- 🔬 QA survey / demo:
|
| 121 |
+
https://tylazhao.github.io/pacific-survey/
|
| 122 |
+
|
| 123 |
+
> Note: Full multi-turn conversations for each preference and expanded usage examples are planned additions to the repo.
|
| 124 |
+
|
| 125 |
+
|
| 126 |
+
Please email tzhao15@uci.edu if you are interested in this project.
|
| 127 |
+
|
| 128 |
+
If you find this work useful, please cite our paper:
|
| 129 |
+
|
| 130 |
+
```bibtex
|
| 131 |
+
@article{zhao2026can,
|
| 132 |
+
title={Can LLMs Discern the Traits Influencing Your Preferences? Evaluating Personality-Driven Preference Alignment in LLMs},
|
| 133 |
+
author={Zhao, Tianyu and Li, Siqi and Shoukry, Yasser and Elmalaki, Salma},
|
| 134 |
+
journal={arXiv preprint arXiv:2602.07181},
|
| 135 |
+
year={2026}
|
| 136 |
+
}
|
| 137 |
+
```
|
| 138 |
+
|
| 139 |
+
|