Master Email Writing with ChatGPT 4o Ensures Professional Communication • Choose AI tone
• Boost email clarity
• Save drafting time Read more: https://
buff.ly/3yVWwcx
LLMS
-
Master Email Writing with ChatGPT 4o Ensures Professional Communication
By
–
-
Discussion on Gemini Live product release status and rollout
By
–
I think it was a plan to launch it today b/c in today's announcement placeholder there were 3 items but only 1 got published (gemini live) but if it didn't go well it may get postponed
-
CEO shares Llama 405B record performance at Global AI Summit
By
–
Our team will be at @globalaisummit 2024 next week where our CEO & Co-Founder, @RodrigoLiang
, will share more about our world record performance on the best open source model, Llama 405B. See you at #GAIN2024! Learn more https://
globalaisummit.org/en/default.aspx #AI #LLM -
Comparing Gemini’s memory and custom system prompt functionality
By
–
Unlike on ChatGPT, you will be able to add your info manually so Gemini can pick it up. This feature now feels to be closer to a custom system prompt rather than Memory
-
New Math Benchmark Developed to Evaluate AI Models
By
–
Given that our Math leaderboard (GSM1K) is now relatively saturated (most models score >90), we are working on producing a new Math benchmark to properly discern between models.
-
Grok 2 Updates Coming in Upcoming Weeks
By
–
Expect more updates, including the addition of Grok 2, in the coming weeks!
-
Claude 3.5 Sonnet and Llama 4.1 405B Excel in Instruction Following
By
–
INSTRUCTION FOLLOWING: Claude 3.5 Sonnet and Llama 4.1 405B Instruct stand out as models with BOTH: – very high (>0.9) Main Request fulfillment
– very high (>0.7) Constraint fulfillment Also, all Claude models take a hit for Structural Clarity in writing style -
Models Show Weaker Instruction Following Performance in Spanish
By
–
MULTILINGUAL (SPANISH) Lastly, one interesting pattern we noticed is that every model seemed to perform worse at Instruction Following (Main Request fulfillment and Constraint fulfillment) in Spanish versus English. This implies there's still meaningful headroom for models to
-
GPT-4 Turbo Remains Most Factual Model Among LLMs
By
–
INSTRUCTION FOLLOWING (FACTUALITY): We find that GPT-4 Turbo is still the most factual model among all of the models we've evaluated. With the trend towards smaller, distilled models, we notice that it seems to come at the cost of performance on factuality.
-
GPT-4o leads code correctness; Claude excels prompt adherence
By
–
Some insights from the leaderboards: CODING We notice that GPT-4o (August 2024) performs the best on Code Correctness among all the models, but Claude outdoes it on Prompt Adherence