SAFETY

–

04 March 2026 21h51

most people building AI agents obsess over how they WRITE memories
turns out that's basically irrelevant new research analyzed 9 different memory systems across 1,540 questions the finding?
retrieval method drives 20-point accuracy swings
write strategy? 3–8 points max raw

→ View original post on X — @testingcatalog

Perplexity’s Final Pass for document analysis

By

@testingcatalog

–

04 March 2026 14h51

Perplexity is working on a new feature called "Final Pass," designed to perform comprehensive document analysis and fact-checking.

Models Self-Regulate Reasoning Length

By

–

04 March 2026 12h19

the deeper point here connects to something the field keeps rediscovering. we trained reasoning models to think longer. then we discovered longer doesn't mean better. now this paper shows the models themselves already know that. they're generating stop signals that our inference

Three Key Observations on Model Paths

By

–

04 March 2026 12h18

three observations back this up: > high-confidence paths (measured by cumulative probability) are shorter AND more accurate than standard samples > at those natural endpoints, the stop token is the model's top prediction. it knows it's finished > as you widen exploration (more

Researchers test AI self-awareness with TSearch

By

–

04 March 2026 12h18

here's where it gets interesting. the researchers probed whether models internally "know" they're done. they introduced TSearch, which scores partial reasoning traces by cumulative log-probability across the entire chain, not just the next token. when you let the model explore

Overthinking hurts accuracy in AI responses

By

–

04 March 2026 12h18

and it's not just wasted compute. overthinking actively hurts accuracy. DeepSeek-R1 produces responses 5x longer than Claude 3.7 Sonnet on AIME 2025 with comparable accuracy. QwQ-32B scores 2 percentage points HIGHER with its shortest answers using 31% fewer tokens. 72% of

RFCS Metric Reveals Early Correct Steps in Models

By

–

04 March 2026 12h18

first, the problem quantified. the researchers created a metric called RFCS (Ratio of First Correct Step) that tracks where in a chain of thought the correct answer first appears. on MATH-500, across every model tested, the right answer shows up well before the end in over half

Overthinking in AI models is a sampling issue

By

–

04 March 2026 12h18

reasoning models already know when they've solved the problem. we just don't let them stop. new paper from Beihang University and ByteDance shows that the overthinking problem in models like DeepSeek-R1 and Qwen3 isn't a training failure. it's a sampling failure. the fix cuts