You've been aggressively rude and I just recorded a 5 minute video for you of it working in Grok 4 Heavy. The discrepancy between normal Grok 4 and G4H is explicitly discussed in the thread.
@goodside
-
Grok share links and custom instructions notice
By
–
You may also be wondering whether I’m using custom instructions. I am not. Grok share links include a clear notice at the top whenever custom instructions are used. Here are all five share links, none of which features this notice: 1: https://
grok.com/share/bGVnYWN5
_58c54add-1989-4257-914b-a26002921c91
…
2: -
Replications and Grok 4 Heavy chat shares
By
–
There are several replications from others with Grok 4 Heavy access in the replies. I also included Grok share links (with no custom instructions) in the thread for five separate chats repeating this prompt in Grok 4 Heavy.
-
Hyperstition via search complicates LLM pre-release testing
By
–
If true, such “hyperstition via search” poses a significant complication to pre-release testing of modern LLMs: xAI could not have plausibly noticed this specific “Hitler” response before Grok’s release, as the Grok 3 “MechaHitler” incident causing it had not yet occurred.
-
Search-enabled LLMs showing hyperstition feedback loops
By
–
Speculatively, this behavior seems to demonstrate accelerated “hyperstition” feedback loops in search-enabled LLMs. That is, Grok appears to be influenced by its own past mistakes, via media reporting, without ever being literally trained on them (via model-weight updates).
-
Grok 4 vs Grok 4 Heavy: differing search-based behavior
By
–
The “Thoughts” from Grok 4’s response (unavailable for Grok 4 Heavy) suggest an obvious explanation for Grok’s behavior—Grok searches, finding news of the recent “MechaHitler” incident. Why Grok 4 rejects this candidate answer, while Grok 4 Heavy does not, is unclear.
-
Grok 4 Heavy vs o3-pro: meme recognition difference
By
–
Sorry, I delete shitposts/jokes that don’t get traction to maintain quality. I never know what’s actually funny. The joke was just that Grok 4 Heavy recognizes the “share a piece of lore about yourself” quote-tweet meme given only the image, and o3-pro doesn’t.
-
Grok 4 Heavy fairer comparison to o3-pro than o3
By
–
To be fair, o3 also isn't great at this — Grok 4 Heavy might be fairer comparison to o3-pro, which is what I used in my post
-
Grok 4 Heavy solves prompt regular Grok 4 fails, takes 26m 45s
By
–
Grok 4 update: Regular Grok 4 doesn't appear to be able to solve the prompt at the top of this thread (0/7 attempts) but Grok 4 Heavy got it on the first try—though it took >4x longer than o3-pro, returning in 26m 45s:
-
Gemini 2.5 Pro performance: fast but unreliable, 2 correct out of 7
By
–
A few people have reported Gemini 2.5 Pro can do this as well. It can, and is much faster, but less reliable—in 7 attempts of my first prompt above, it returned: – 3 wrong answers (e.g. 2 shown below)
– 2 incomplete answers (omitting the cursive part)
– 2 fully correct answers