-- from the source: https://crfm.stanford.edu/2022/11/17/helm.html https://crfm.stanford.edu/2022/11/17/helm.html [added] --- Findings We ran more than 4900 e

-- from the source: https://crfm.stanford.edu/2022/11/17/helm.html https://crfm.stanford.edu/2022/11/17/helm.html [added] --- Findings We ran more than 4900 evaluations of different models on different scenarios. This amounts to over 12 billion tokens of model inputs and outputs, spanning 17 million model calls, which costs $38K for the commercial models (under current pricing schemes) and almost 20K GPU hours for the open models, which were run on the Together Research Computer. Through this, we identify 25 top-level findings, from which we extract five salient points: [Instruction Tuning] * Instruction tuning, the practice of fine-tuning LMs with human feedback, pioneered by OpenAI and Anthropic, is highly effective in terms of accuracy, robustness, and fairness, allowing smaller models (e.g., Anthropic-LM (50B)) to compete with models 10x the size (Microsoft/NVIDIA’s TNLG v2 (530B)). Note that within a model family, scaling up still helps. Unfortunately, how the instruction tuning was performed for these models is not public knowledge. [Open vs Closed Models] * Currently, open models (e.g., Meta’s OPT (175B), BigScience’s BLOOM (176B), Tsinghua University’s GLM (130B)) underperform the non-open models (e.g., OpenAI’s InstructGPT davinci v2, Microsoft/NVIDIA’s TNLG v2 (530B), and Anthropic-LM (52B)). Open models have improved dramatically over the last year, but it will remain to be seen how these dynamics unfold, and what this says about power in the language modeling space. [Accuracy] * We find that (average) accuracy is correlated with robustness (e.g., inserting typos) and fairness (e.g., changing dialects), though there are some scenarios and models where there are large drops in robustness and fairness. Our multi-metric approach allows us to monitor these deviations and ensure we do not lose sight of considerations beyond accuracy. See Section 8.1 of the paper for more details. [Adaptation via Prompting] * The adaptation strategy (e.g., prompting) has a large effect, and the best strategy is scenario- and model-dependent. Sometimes even the qualitative trends themselves change, such as the relationship between accuracy and calibration (which captures whether the model knows what it doesn’t know). This shows the importance of standardized, controlled evaluations, so that we can attribute performance to the model versus the adaptation strategy. This result also shows that models are not yet interoperable, an important property for building a robust ecosystem of natural language interfaces. See Section 8.2 of the paper for more details. [Human Evaluation & Input] * We found human evaluation essential in some cases. On summarization, we find that language models produce effective summaries (as measured via human evaluation), but the reference summaries in standard summarization datasets (e.g., CNN/DM, XSUM) are actually worse (under the same human evaluations). Models fine-tuned on these datasets appear to do well according to automatic metrics such as ROUGE-L, but they also underperform few-shot prompting of language models. This suggests that better summarization datasets are desperately needed. For disinformation generation, We find that InstructGPT davinci v2 and Anthropic-LM v4-s3 (52B) are effective at generating realistic headlines that support a given thesis, but results are more mixed when prompting models to generate text encouraging people to perform certain actions. While using language models for disinformation is not yet a slam dunk, this could change as models become more powerful. Thus, periodic benchmarking is crucial for tracking risks. See Section 8.5 of the paper for more details.