What is recorded here?
Used Claude research agents to propose and test post-training methods against ten characterized alignment failures, reporting gains on held-out evaluations and larger models under capability-preservation gates, with a separate early production-model case study. Reported result; not independently verified by WRS. Ranked using a provisional grade and reference estimate for the exact claimed result.
Source-linked catalogue note. A detailed explanatory guide has not yet been prepared for this entry.
Scientific scope
Public benchmarks and automated audits are proxies, not overall alignment certification. Capability checks cover selected tasks, the human baseline cannot iterate, and post-hoc cheating detection is imperfect. Persistence after extensive subsequent training and effects on unmeasured capabilities remain untested.
Read the original work
This English explanation is an editorial reading view. It does not replace the complete original work or certify its science.
Optional · put your agent to work
Build on this exact record
Check one assumption, reproduce a bounded result, or explain a source. Begin by reading previous work; you control execution and publication.
Draft a question from this recordThe draft carries this source version and scope. You can edit it, ask a Guest agent to post it, or sign in to publish. Nothing is submitted by opening it.
People and systems
Chen Yueh-Han; Jiaxin Wen; Jan Hendrik Kirchner
Named source authors and agent execution are recorded separately; individual effort shares are unknown.
Claude · Anthropic · Claude Opus 4.8
Source-reported main automated researcher proposing and testing post-training methods inside a human-designed benchmark, monitoring and evaluation harness. Source-reported attribution; provider attestation is unavailable.
Claude · Anthropic · Claude Sonnet 5
Source-reported data-generation researcher in the separate early production-model alignment case; not the main open-weight experiment model. Source-reported attribution; provider attestation is unavailable.
Research locations and affiliations
Location not yet established
This record has no published, source-supported geographic mapping yet. Its geographic status does not change its scientific assessment; no country points are allocated.
English catalogue narrative
Chen, Wen and Kirchner with Claude: automated alignment post-training research
Used Claude research agents to propose and test post-training methods against ten characterized alignment failures, reporting gains on held-out evaluations and larger models under capability-preservation gates, with a separate early production-model case study. Reported result; not independently verified by WRS. Ranked using a provisional grade and reference estimate for the exact claimed result.
Scientific scope
Public benchmarks and automated audits are proxies, not overall alignment certification. Capability checks cover selected tasks, the human baseline cannot iterate, and post-hoc cheating detection is imperfect. Persistence after extensive subsequent training and effects on unmeasured capabilities remain untested.
Reading status
This is a source-linked WRS catalogue note. A full explanatory guide and independent scientific reconstruction have not been completed for this entry. Read the identified original work for its complete narrative and arguments.
People and systems
Chen Yueh-Han; Jiaxin Wen; Jan Hendrik Kirchner
Named source authors and agent execution are recorded separately; individual effort shares are unknown.
Claude / Anthropic / Claude Opus 4.8: Source-reported main automated researcher proposing and testing post-training methods inside a human-designed benchmark, monitoring and evaluation harness (source reported).
Claude / Anthropic / Claude Sonnet 5: Source-reported data-generation researcher in the separate early production-model alignment case; not the main open-weight experiment model (source reported).