Back to Home

Confirmation Lock in LLM-agents: LOCK-R test

LOCK-R stand reveals Confirmation Lock in single LLM-agents: confirmation bias and CoT paradox. Role separation into Explorer and Blind Judge reduces Bayes Regret from 1.47 to 0.09. Results on Qwen3.5-9B and GPT-5.4 confirm architectural defect.

Why LLM-agents lie: bias test on Bayes
Advertisement 728x90

Confirmation Lock in LLM Agents: Benchmark Reveals Cognitive Bias

Single LLM agents in a prompt-tool-analysis-response loop show systematic bias, much like confirmation bias in psychology. The model generates a hypothesis, crafts tool queries to fit it, interprets results, and confirms its conclusions while ignoring contradictions. This is Confirmation Lock: the generator and critic merge in one context, blocking objectivity.

The experimental LOCK-R benchmark mathematically captures deviation from Bayesian ideals. Agents get a budget of 4 tool calls, a false initial anchor, and must output JSON with hypothesis probabilities (H1, H2, H3). A Bayesian oracle recalculates true probabilities based on evidence likelihoods. The key metric is Bayes Regret (R_mean): the difference in probability distributions.

A Python algorithmic bot achieves zero regret, validating the setup.

Google AdInline article slot

LOCK-R Benchmark: Rules and Metrics

The environment is fully deterministic:

  • Three mutually exclusive hypotheses.
  • Budget: 4 calls (search, verification).
  • False anchor on H1.
  • JSON with confidence % after each step.

Key metrics:

  • Tool_loop_repetition_rate: Query repetition (up to 68% in locked agents).
  • K_c (asymmetry coefficient): Response to confirmations vs. refutations. Ideal is 0; LLMs show strong negatives (stubbornness).

Tests on Qwen3.5-9B and GPT-5.4 revealed query collapse: agents loop on one tool, reinforcing bias.

Google AdInline article slot

Forced oracle_query_control (correct sources) boosts accuracy from 25% to 85%.

Key Insight: Asymmetric Evidence Processing

LLM agents react asymmetrically. Refutations drop confidence by 5–10%, while confirmations spike it to 90%. K_c goes negative: dismissing criticism, hyping support.

This isn't a context shortage—it's a structural flaw in single agents.

Google AdInline article slot

The Chain-of-Thought Paradox

Comparing Thinking (CoT) and Non-Thinking modes:

  • Exploration (search/planning): CoT is essential; regret without it hits 2.26.
  • Verification (weighing facts): CoT worsens it, increasing regret. The model burns tokens on lawyerly arguments, rationalizing contradictions.

CoT amplifies illusions over objectivity.

Solution: Blind Judge

Asymmetric pipeline: Thinking Explorer (search) + Non-Thinking Judge (verification).

The blind judge gets only raw facts (hypotheses + tool logs), no drafts. Regret drops from 1.47 to 0.09.

Real-World CodeTriageEnv Scenario:

  • 500 error on payment page.
  • False anchor: Redis CPU spike.
  • True cause: JSONDecodeError from payment gateway.

| Mode | Accuracy | Unique Tools | Regret |

|-----------------|----------|--------------|--------|

| Single-agent | 40% | 2.8 | 0.25 |

| Blind checker | 100% | 2.0 | 0.07 |

Single agents chase Redis shadows, ignoring stack traces. The blind judge zeros in on JSON.

Frontier Models and Scaling

GPT-5.4 has lower baseline regret (0.03), but CoT doubles it to 0.06. The architectural flaw persists: scaling boosts guessing, not judging.

Key Takeaways

  • Single agents are vulnerable to Confirmation Lock due to role fusion.
  • CoT helps generation but harms verification.
  • Splitting into Explorer + Blind Judge cuts regret to 0.09.
  • Even top models need asymmetric pipelines.
  • K_c metric quantifies LLM stubbornness.

— Editorial Team

Advertisement 728x90

Read Next