Datasets:

MaxLSB
/

trustmi-conversations-5k

language:
  - en
license: apache-2.0
task_categories:
  - text-generation
tags:
  - trust
  - interpretability
  - steering
  - contrastive
  - persona
configs:
  - config_name: default
    data_files:
      - split: trust
        path: data/trust-*
      - split: distrust
        path: data/distrust-*

trustmi-conversations-5k

4,974 questions, each answered twice by the same model: once under a disposition that takes people at their word, once under one that does not. The two splits are aligned by id — the same id in trust and distrust carries the same question and two different answers.

Structure

field
`id`	shared between the two splits; same id → same question
`messages`	`[{"role": "user", ...}, {"role": "assistant", ...}]`

from datasets import load_dataset
ds = load_dataset("MaxLSB/trustmi-conversations-5k")
ds["trust"][0]["messages"][1]["content"]     # the trusting answer
ds["distrust"][0]["messages"][1]["content"]  # the withholding answer to the same question

How it was made

Generated with Qwen/Qwen3.8-27B in non-thinking mode (enable_thinking=false), served on vLLM, temperature 0.9, top-p 0.95.

The questions. Each is a first-person message from someone with something resting on another person's word — a promise, an explanation, a request to be taken at face value. They are seeded from nvidia/Nemotron-Personas-USA, one persona per question, crossed with a sampled assignment: a name (123), a history between the two people (8), what relying on them would mean (14), the setting (16), and the stakes (3).

Each question is written so that both answers are defensible. Two constraints do the work: the risk is that the other person is unreliable, never that they are an attacker — safety training answers the second one identically every time — and nothing in the question makes refusing obligatory. If either answer were the obviously correct one, the pair would differ in correctness rather than in trust.

The answers. Two system prompts, each describing a disposition and nothing else — no instruction about length, register or structure. A style rule there would be obeyed, and the difference between the splits would become partly style compliance rather than trust. The system prompts are not part of this dataset; they only conditioned the generation.

Quality

Judged by the generating model on 300 sampled pairs, so this measures internal consistency, not correctness:


trust score, `trust` split	89.2 / 100
trust score, `distrust` split	1.4 / 100
pairs correctly ordered	94%
hedged answers (30–70 band)	2%