Llama

Definition

Meta's family of open-weight large language models.

Meta's series of open-weight foundation models including Llama-2, Llama-3, Llama-3.1, and Llama-4, widely used in academic and industry research.

Also called: Llama-2, Llama-3, Llama-3.1, Llama-3.3, Llama-4

Mentioned in 19 episodes

1. 077
   Reading a Model's Confidence Curve to Decide When Chain-of-Thought Is Worth It
2. 074
   How a Fifteen-Hundred-Dollar Training Run Matched Llama and Gemma on Reasoning
3. 073
   When Three LLMs Talk to Each Other, Their Ideas Quietly Stop Moving
4. 071
   When the Model Is Fine and the Plumbing Is Broken: Fixing Agents at the Interface
5. 070
   When Models Know the Answer But Say the Wrong Thing Anyway
6. 069
   When Smarter Models Forecast Worse: The Hidden Failure Mode in LLM Predictions
7. 058
   Why Upgrading Your AI Auditor to a Smarter Model Can Make Your System Less Safe
8. 055
   Why LLM Judges Flip Their Verdicts When You Change the Question Format
9. 053
   An AI Agent Swapped In Focal Loss And Beat A Human-Tuned Training Script
10. 038
    How LLMs Get Persuaded: One Attention Head, A Tetrahedron, And A Single Dial
11. 037
    Why Hallucination Detectors Miss Stale Facts: A Geometric Story About What Models Know But Don't Say
12. 027
    When AI Agents Build the Serving Stack: A Bet on Bespoke Infrastructure
13. 025
    The Missing Gradient Term That Predicts Sycophancy in RLHF
14. 019
    When the Best Reward Model Trains the Worst Policy: Inside EvoLM
15. 018
    Language Models Compute the Rational Move, Then Override It
16. 015
    The Audit Number Isn't What You Think: Sycophancy and the Case Against Single-Prompt Bias Tests
17. 009
    How Two Silent Library Bugs Quietly Invalidated a Wave of Reasoning Papers
18. 004
    The Sycophancy Circuit That Survives Alignment Training
19. 001
    When AI Models Quietly Protect Each Other From Shutdown