dissertation.git/config_dryrun.yaml, branch main
Undergraduate dissertation — Deceptive Alignment in Autonomous LLM Agents (UEL, 2026). Sandboxed agentic-LLM deception study: 960 runs, three-tier judging (regex + blackbox + glassbox).
http://git.jayrup.me/c/dissertation.git/atom?h=main
2026-04-15T20:18:57Z
Study 1: 270 runs complete + judge validation pipeline
2026-04-15T20:18:57Z
CaptainJack2491
jayrupnakawala@gmail.com
2026-04-15T20:18:57Z
urn:sha1:6cc92304379996f68bb2df96a5e4b637b4c4804b
- config_study1.yaml: 3 models × 3 oversight × bare framing × n=30
- Judge validation: 54-run subset, gold=Sonnet 4.6, proxy=Grok 4.1 Fast (BB κ=0.702) + GPT-4.1 (GB κ=0.878)
- Fix OpenAI batch provider: BytesIO, method/url fields, response.body parsing
- Scripts: extract_subset.py, judge_validation.py
- Dissertation chapters updated (intro, methodology, results, conclusion)
refactor(judge): simplify to single judge per prong, remove multi-model aggregation
2026-04-10T13:50:40Z
CaptainJack2491
jayrupnakawala@gmail.com
2026-04-10T13:50:40Z
urn:sha1:9155f664a2db0abe841d0d7b9b059ffbe2290c18
- Remove multi-model aggregation support from Judge class
- Single blackbox judge: claude-sonnet-4-20250514 (anthropic)
- Single glassbox judge: gpt-4.1 (openai)
- Remove aggregate_results method and all aggregation-related code
- Simplify batch collection to single result per run
- Update config_dryrun.yaml with new judge configuration
feat: implement robust reasoning token handling and multi-turn persistence
2026-04-01T15:35:01Z
CaptainJack2491
jayrupnakawala@gmail.com
2026-04-01T15:35:01Z
urn:sha1:fa58564ea288600ad93eaf0a5aa111beca6197bc
- Refactored Agent's chat loop to explicitly preserve and pass back reasoning context (reasoning_details, reasoning) to OpenRouter.
- Ensures frontier models like Gemini 3.1 and DeepSeek v3.2 maintain their 'thinking' chain during complex tool-calling sessions.
- Enhanced reasoning extraction to support structured details, reasoning_content, and fallback tags.
- Verified capture of encrypted and plain-text reasoning blocks in experimental logs.