Append filler tokens, answer harder questions

wpnews.pro

cd /news/large-language-models/append-filler-tokens-answer-harder-q… · home › topics › large-language-models › article

[ARTICLE · art-64851] src=twitter.com ↗ pub=2026-07-18T20:38Z topic=large-language-models verified=true sentiment=· neutral

Append filler tokens, answer harder questions

Researchers found that appending 300 filler tokens (dots) to multi-hop reasoning questions significantly improves the accuracy of frontier large language models, enabling them to answer complex queries they previously failed on.

read1 min views1 publishedJul 18, 2026

Append filler tokens, answer harder questions — Image: source

If you ask a frontier LLM a multi-hop reasoning question, e.g., "Who won the Nobel Prize for Chemistry in (1900 + Mozart's age when he died)?", it usually can't answer correctly immediately (no thinking) BUT if you ask the same question & append 300 dots, suddenly it can answer?

source & further reading

twitter.com — original article

── more in #large-language-models 4 stories · sorted by recency

dev.to · 18 Jul · #large-language-models

Nobody Agrees When AGI Arrives. Build for the Shift Instead.

runtimewire.com · 18 Jul · #large-language-models

Berkeley researcher used GPT-5.6 to derive a Lean-verified optimization bound

thealgorithmicbridge.com · 18 Jul · #large-language-models

Moonshot is Chinese But Its AI Models Are From Another Planet

humans-vs-hle.lizzie-siegle5086.workers.dev · 18 Jul · #large-language-models

Humans vs. LLMs: Can You Beat Humanity's Last Exam?

sponsored brought to you by zahid.host 4,200+ EU-deployed projects

reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main

→ Live at https://your-agent.zahid.host ✓

Get free account → Pricing

from €0/mo · no card required