LLM Digest
Subscribe

AI Storyline

11 items · 5 sources · 7 days

View as JSON

Operational story trace

Empirical Audit

Latest change

On Nov 16 the audits reached security: TRIAGE-CPG compares several LLMs on budgeted static malware triage, testing which models can clear, flag or escalate unknown binaries under a fixed analysis budget.

Earlier contextThe story so far

This is a research trend, not one developing event: eleven unrelated arXiv papers from Oct 30 to Nov 16 that each re-test a technique already in use. It began with audits of inference cost (early stopping, test-time search, small-model compression) and spread to alignment, token-spend metrics, KV-cache reuse, benchmark integrity and agent debugging.

editor-curated · source-linked

Arc

Oct 30Nov 16 · now
COMPUTE · Oct 30
Audits of inference cost: early stopping, test-time search and compression
3 sources · show sources ▾
ALIGNMENT · Nov 3
Direct Nash Optimization tested as a practical alignment recipe
1 source · show source ▾
COST & INFRA · Nov 10
Token spend as a productivity metric, and KV caches across model updates
2 sources · show sources ▾
EVAL INTEGRITY · Nov 11–13
Benchmark scores can be throttled, and persona steering has safety side effects
Two papers question signals teams use to judge model capability and safety.
2 sources · show sources ▾
AGENT FAILURES · Nov 9–13
Where errors enter long agent trajectories, and whether interpretability can explain them
2 sources · show sources ▾
NOW · Nov 16
Cross-model audit of LLMs for budgeted malware triage
1 source · show source ▾

What to watch — open questions

  • Do the early-stopping and test-time search audits publish thresholds a serving stack can adopt directly?
  • Does the compression ablation name a quantization-plus-pruning order that is safe to combine on small models?
  • Does the DeepSeek-V4-Flash KV-cache result hold for other model families or larger weight updates?
  • Does any provider or eval harness add checks for deliberate score throttling (sandbagging)?
  • Does the trajectory-centered error study release its traces or controls so other agent teams can reproduce it?
  • Does TRIAGE-CPG release its benchmark so security teams can rerun the model comparison on their own samples?
How this thread was built
editor wrote the arc · 6 beats

Storylines are threaded mechanically from the feed: stories that share a distinctive anchor across multiple days and sources. Each item links to its original source. The evidence trace, current state, and open questions are written by the editor routine and refreshed whenever a new beat lands.