cd /news/large-language-models/completely-inaccurate-results-from-f… · home topics large-language-models article
[ARTICLE · art-28488] src=discuss.huggingface.co ↗ pub= topic=large-language-models verified=true sentiment=↓ negative

Completely inaccurate results from file read

A user reports that LM Studio's RAG system fails to read uploaded PDF files, instead generating inaccurate results based on internet searches. The issue persists across multiple models including DeepSeek, Qwen3-14b, and Llama 3 8b, producing poor outputs for tasks like creating discussion worksheets and blurbs.

read1 min views1 publishedJun 15, 2026

Hello, I am new to all this. I have LM Studio set up and downloaded several models. I also set up RAG (I think). I am trying to use it to read a pdf of one of my novels and create a discussion worksheet, a new blurb, etc.

I have tried a few different models to do this, including DeepSeek, Qwen3-14b and Llama 3 8b. None of them seem to actually read the pdf file I provide. At best, it seems to search the internet for details and create something from that. The results have been poor to absolute rubbish based on whatever it seems to be getting the information from.

I have very little idea of what I am doing here, so if there is a better approach, any help is greatly appreciated.

── more in #large-language-models 4 stories · sorted by recency
── more on @lm studio 3 stories trending now
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
LIVE [news/completely-inaccurat…] indexed:0 read:1min 2026-06-15 ·