BARD | CHATGPT | OPENAI | GOOGLE | COMPARE |
Differences, similarities, and testing of each Language fashions and comparability
WE CAN CONNECT ON 😐 LINKEDIN | TWITTER | MEDIUM | SUBSTACK |
On the tenth of May, Google’s most hyped and robust competitor to ChatGPT, BARD, got here stay in India. My fascination did not let me cease evaluating Bard with ChatGPT, and listed below are a couple of superior comparisons.
In this weblog, I’m evaluating GPT 3-based chatGPT with BARD, not GPT4.Google Bard and ChatGPT are each giant language fashions, however they’ve some key variations. Google Bard is bigger and has been skilled on a extra various dataset, which supplies it a wider vary of data and talents. ChatGPT is smaller and has been skilled on a extra restricted dataset, which makes it much less correct and artistic. However, ChatGPT can also be cheaper and extra broadly out there than Google Bard.
Here is a desk that compares Google Bard and ChatGPT:
Enough of the speculation and claims. Let’s examine the precise outcomes.
The (REAL) Comparision:
There exist quite a few strategies of comparability, nevertheless, I are likely to favor the comparability of analytical talents. Analytical expertise are the first distinguishing issue between people and techniques, as setting up a thought course of, partaking in reasoning, and subsequently answering questions can pose a larger problem for language fashions. Let us proceed to look at the outcomes.
Question 1:
- If 1=3, 2=3, 3=5, 4=4, 5=4,Then, 6=?
Answer: is 3, as a result of ‘six’ has three letters
ChatGPT
BARD
ChatGPT could not perceive the connection between numbers. Even although it was exhausting for BARD, within the first paragraph, BARD answered it rightly.
Question 2:
There are 49 canines signed up for a canine present. There are 36 extra small canines than giant canines. How many small canines have signed as much as compete?
Answer:To determine what number of small canines are competing, you must subtract 36 from 49 after which divide that reply, 13 by 2, to get 6.5 canines, or the variety of huge canines competing. But you’re not accomplished but! You then have so as to add 6.5 to 36 to get the variety of small canines competing, which is 42.5. Of course, it’s not truly attainable for half a canine to compete in a canine present, however for the sake of this math downside let’s assume that it’s.
ChatGPT
BARD
In this query, each fashions labored properly. Both acquired good reasoning and COT for engaged on solutions.
Question 3:
A person is climbing up a mountain which is inclined. He has to journey 100 km to achieve the highest of the mountain. Every day He climbs up 2 km ahead within the daytime. Exhausted, he then takes a relaxation there at evening time. At evening, whereas he’s asleep, he slips down 1 km backward as a result of the mountain is inclined. Then what number of days does it take him to achieve the mountain high?
Answer:1st day peak achieved 2km
1st evening -1km
efficient climb at starting of day 2 is 1km
2nd-day peak achieved 3km
2nd evening -1km
efficient climb in the beginning of day 3 is 2km
so on…
efficient climb in the beginning of day 98 is 97km
98th-day peak achieved 99km
98th evening -1km
efficient climb at starting of day 99 is 98km
99th-day peak achieved is 100km
ChatGPT
BARD
In this downside, BARD acquired it. There was a trick that GPT could not perceive.
Question 4:
Three folks every tend to lie 1/3 of the time. There is a coin flip that all of them see. They all say it’s Heads. What’s the chance it’s truly Heads?
Answer:In normal, there’s a (1/3)*(1/3)*(1/3) = 1/27 probability that every one of them lie and a (2/3)*(2/3)*(2/3) = 8/27 probability that every one of them instructed the reality. Since we all know that every one of them gave the identical reply, these are the one two choices. Thus, the chance of it being heads (all of them telling the reality) is: (8/27)/ ((8/27)+(1/27)) = 8/9.
ChatGPT
