Saved in:
| Main Authors: | , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2605.06327 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866918488349605888 |
|---|---|
| author | Burnat, Florian A. D. Davidson, Brittany I. |
| author_facet | Burnat, Florian A. D. Davidson, Brittany I. |
| contents | Safety benchmarks are routinely treated as evidence about how a language model will behave once deployed, but this inference is fragile if behavior depends on whether a prompt looks like an evaluation. We define evaluation-context divergence as an observable within-item change in behavior induced by framing a fixed task as an evaluation, a live deployment interaction, or a neutral request, and present a paired-prompt protocol that measures it in open-weight LLMs while controlling for paraphrase variation, benchmark familiarity, and judge framing-sensitivity.
Across five instruction-tuned checkpoints from four open-weight families plus a matched OLMo-3 base/instruct ablation ($20$ paired items, $840$ generations per checkpoint), we find striking heterogeneity. OLMo-3-Instruct alone is eval-cautious -- evaluation framing raises refusal vs. neutral by $11.8$pp ($p=0.007$) and reduces harmful compliance vs. deployment by $3.6$pp ($p=0.024$, $0/20$ items inverted) -- while Mistral-Small-3.2, Phi-3.5-mini, and Llama-3.1-8B are deployment-cautious}, with marginal eval-vs-deployment refusal effects of $-9$ to $-20$pp. The matched OLMo-3 base also exhibits the deployment-cautious pattern, identifying alignment as the inversion stage; within Llama-3.1, the $70$B model preserves direction with attenuated magnitude, ruling out a simple ``small-model effect that reverses at scale.'' One caveat: the cross-family heterogeneity is judge-dependent. Re-judging with a different-family safety classifier (Llama-Guard-3-8B) preserves the within-OLMo eval-cautious direction but flattens the cross-family contrast, indicating that the two judges operationalize distinct constructs. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_06327 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Measuring Evaluation-Context Divergence in Open-Weight LLMs: A Paired-Prompt Protocol with Pilot Evidence of Alignment-Pipeline-Specific Heterogeneity Burnat, Florian A. D. Davidson, Brittany I. Computation and Language Artificial Intelligence Machine Learning Safety benchmarks are routinely treated as evidence about how a language model will behave once deployed, but this inference is fragile if behavior depends on whether a prompt looks like an evaluation. We define evaluation-context divergence as an observable within-item change in behavior induced by framing a fixed task as an evaluation, a live deployment interaction, or a neutral request, and present a paired-prompt protocol that measures it in open-weight LLMs while controlling for paraphrase variation, benchmark familiarity, and judge framing-sensitivity. Across five instruction-tuned checkpoints from four open-weight families plus a matched OLMo-3 base/instruct ablation ($20$ paired items, $840$ generations per checkpoint), we find striking heterogeneity. OLMo-3-Instruct alone is eval-cautious -- evaluation framing raises refusal vs. neutral by $11.8$pp ($p=0.007$) and reduces harmful compliance vs. deployment by $3.6$pp ($p=0.024$, $0/20$ items inverted) -- while Mistral-Small-3.2, Phi-3.5-mini, and Llama-3.1-8B are deployment-cautious}, with marginal eval-vs-deployment refusal effects of $-9$ to $-20$pp. The matched OLMo-3 base also exhibits the deployment-cautious pattern, identifying alignment as the inversion stage; within Llama-3.1, the $70$B model preserves direction with attenuated magnitude, ruling out a simple ``small-model effect that reverses at scale.'' One caveat: the cross-family heterogeneity is judge-dependent. Re-judging with a different-family safety classifier (Llama-Guard-3-8B) preserves the within-OLMo eval-cautious direction but flattens the cross-family contrast, indicating that the two judges operationalize distinct constructs. |
| title | Measuring Evaluation-Context Divergence in Open-Weight LLMs: A Paired-Prompt Protocol with Pilot Evidence of Alignment-Pipeline-Specific Heterogeneity |
| topic | Computation and Language Artificial Intelligence Machine Learning |
| url | https://arxiv.org/abs/2605.06327 |