This is <i>almost</i> perfect.<p>The gold standard for LLM evaluation would have the following qualities:<p>1. Categorized (e.g. coding, reasoning, general knowledge)<p>2. Multimodal (at least text and image)<p>3. Multiple difficulties (something like "GPT-4 saturates or scores >90%", a la MMLU, "GPT-4 scores 20-80%", and "GPT-4 scores < 10%")<p>4. Hidden (under 10% of the dataset publicly available, enough methodological detail to inspire confidence but not enough to design to the test set)<p>The standard model card suite with MMLU, HumanEval etc. has already been optimized to the point of diminishing value - Goodhart's law in action. Meanwhile, arena Elo (<a href="https://huggingface.co/spaces/lmsys/chatbot-arena-leaderboard" rel="nofollow">https://huggingface.co/spaces/lmsys/chatbot-arena-leaderboar...</a>) is extremely useful, but also has the drawback of reflecting median-voter preferences that will not necessarily correlate with true intelligence as capabilities continue to advance, in the same sense as how the doctor with the best bedside manner is not necessarily the best doctor.<p>Until that happens, I'll pay attention to every eval I can find, but am also stuck asking "how many r's are in strawberry?" and "draw a 7-sided stop sign" to get a general impression of intelligence independent of gameable or overly general benchmarks.<p>But all that aside:<p><pre><code> Model | Score
</code></pre>
----------------------------------------------<p><pre><code> GPT-4o | 52
Llama 3.1 405B | 50
Claude 3.5 Sonnet | 46
Mistral Large | 44
Gemini 1.5 Pro | 12
</code></pre>
What an incredible contrast to MMLU, where all of these models score in the 80-90% range! For what it's worth, these scores also fall much closer to my impressions from daily use. Gemini is awful, Sonnet and 4o are amazing, and the new Llama puts fine-tunable, open-source 4o in the hands of anyone with a mini-cluster.
Having very little familiarity of models other than using ChatGPT+, I can't quite make sense of how it scores GPT 3.5, GPT 4, and GPT 4o. Going from 3.5 to 4 felt like a revolutionary step up to me, and I still haven't quite grasped the quality improvement between 4 and 4o - 4o just seems a little chattier to me than 4. (My perceptions are through the web, not through the API.)
Would be nice if they at least published the system prompts used (they can affect the performance of the models), and whether they use few-shot prompting or not.