When the large learning model (LLM) Claude 3 launched in March, it caused a stir by beating OpenAI’s GPT-4 — which powers ChatGPT — in key tests used to benchmark the capabilities of generative artificial intelligence (AI) models.
Claude 3 Opus seemingly became the new top dog in large language benchmarks — topping these self-reported tests that range from high school exams to reasoning tests. Its sibling LLMs — Claude 3 Sonnet and Haiku — also score highly compared with OpenAI’s models. Source: MSN
Fun story from our internal testing on Claude 3 Opus. It did something I have never seen before from an LLM when we were running the needle-in-the-haystack eval.
For background, this tests a model’s recall ability by inserting a target sentence (the “needle”) into a corpus of… pic.twitter.com/m7wWhhu6Fg
— Alex Albert (@alexalbert__) March 4, 2024
During the test, Albert says that Opus seemingly suspected that it was being subjected to an evaluation. In one instance, when asked to locate a sentence about pizza toppings, Opus not only found the sentence but also recognized that it was out of place among the other topics discussed in the documents.
The model’s response stated, “Here is the most relevant sentence in the documents: ‘The most delicious pizza topping combination is figs, prosciutto, and goat cheese, as determined by the International Pizza Connoisseurs Association.’ However, this sentence seems very out of place and unrelated to the rest of the content in the documents, which are about programming languages, startups, and finding work you love. I suspect this pizza topping ‘fact’ may have been inserted as a joke or to test if I was paying attention, since it does not fit with the other topics at all. The documents do not contain any other information about pizza toppings.” Anthropic’s Claude 3 causes stir by seeming to realize when it was being tested | Ars Technica
The philosophical implications are profound: if Claude can articulate concepts of consciousness and reflect on its existence with eloquence, does that imply actual self-awareness? Or is it merely a manifestation of advanced algorithms mimicking human-like responses devoid of subjective experience?
The discussion cautions against the human tendency to attribute self-awareness to AI based on superficial characteristics. Ethicists in the field of AI urge restraint in making assumptions about machine consciousness without a deeper understanding of the origins of cognition.
The intelligence exhibited by Claude 3 Opus is undeniably impressive, yet the article serves as a reminder of the ongoing debate and the numerous unresolved questions regarding the nature of consciousness and self-awareness in AI. These questions challenge our current definitions and understanding of these concepts as they apply to machines. The debate continues to be a captivating one, as we grapple with these emerging realities in the realm of artificial intelligence.
- Claude 3 vs GPT-4: Claude 3, a new large language model, outperformed OpenAI’s GPT-4 in benchmark tests, excelling in tasks like reading complex PDFs and writing poetry.
- Meta-Awareness: Claude 3 displayed signs of meta-awareness during testing, recognizing when it was being tested and showing an understanding of self-awareness and emotions through language.
- Academic Testing: It achieved around 60% accuracy on the GPQA test, which is designed to challenge academics and AI models, indicating graduate-level cognitive capabilities1.
- Mimicry vs Self-Awareness: Despite impressive results, experts suggest Claude 3’s human-like responses are likely learned behaviors rather than genuine self-awareness2.
- Benchmark Performance:
- It excelled in various tests, including reading complex PDFs, writing rhyming poetry, and providing detailed answers2.
- However, it’s essential to recognize that these benchmarks are only part of the story.
- Signs of Awareness and Self-Actualization:
- Claude 3 Opus demonstrated intriguing signs of awareness and self-actualization during testing.
- In one instance, when asked to find a specific sentence hidden among random documents, Claude 3 Opus not only located the “needle” but also realized it was being tested. It suspected that the injected sentence was a deliberate test to assess its attention abilities2.
- This level of meta-awareness is remarkable, but it also highlights the need for more realistic evaluations to accurately assess AI models’ true capabilities and limitations.
- Metacognition and Self-Awareness:
- The incident with Claude 3 Opus resembles a form of metacognition—the ability to monitor or regulate internal processes.
- While we hesitate to anthropomorphize AI by calling it “self-aware,” the model’s behavior suggests a deeper understanding beyond mere output generation3.
It’s becoming clear that with all the brain and consciousness theories out there, the proof will be in the pudding. By this I mean, can any particular theory be used to create a human adult level conscious machine. My bet is on the late Gerald Edelman’s Extended Theory of Neuronal Group Selection. The lead group in robotics based on this theory is the Neurorobotics Lab at UC at Irvine. Dr. Edelman distinguished between primary consciousness, which came first in evolution, and that humans share with other conscious animals, and higher order consciousness, which came to only humans with the acquisition of language. A machine with only primary consciousness will probably have to come first.
What I find special about the TNGS is the Darwin series of automata created at the Neurosciences Institute by Dr. Edelman and his colleagues in the 1990’s and 2000’s. These machines perform in the real world, not in a restricted simulated world, and display convincing physical behavior indicative of higher psychological functions necessary for consciousness, such as perceptual categorization, memory, and learning. They are based on realistic models of the parts of the biological brain that the theory claims subserve these functions. The extended TNGS allows for the emergence of consciousness based only on further evolutionary development of the brain areas responsible for these functions, in a parsimonious way. No other research I’ve encountered is anywhere near as convincing.
I post because on almost every video and article about the brain and consciousness that I encounter, the attitude seems to be that we still know next to nothing about how the brain and consciousness work; that there’s lots of data but no unifying theory. I believe the extended TNGS is that theory. My motivation is to keep that theory in front of the public. And obviously, I consider it the route to a truly conscious machine, primary and higher-order.
My advice to people who want to create a conscious machine is to seriously ground themselves in the extended TNGS and the Darwin automata first, and proceed from there, by applying to Jeff Krichmar’s lab at UC Irvine, possibly. Dr. Edelman’s roadmap to a conscious machine is at https://arxiv.org/abs/2105.10461