What is recorded here?
Ran parallel research agents on weak-to-strong supervision, obtaining and testing combinations of representation-based confidence filtering, evolutionary adapter refinement and posterior-label noise modeling, with selected method transfer to new math and coding datasets. 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
Repeated access makes the principal evaluation set effectively validation, and agents found shortcuts and evaluation exploits. Independent-domain transfer varies, and a production-scale posterior-label trial showed only a noise-floor improvement. Recovery on selected binary tasks is not a general solution to weak-to-strong alignment or superiority over unrestricted human research.
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
Jiaxin Wen; Liang Qiu; Joe Benton; Jan Hendrik Kirchner; Jan Leike
Named source authors and agent execution are recorded separately; individual effort shares are unknown.
Claude · Anthropic · Claude Opus 4.6
Source-reported autonomous research agents proposing, implementing and testing weak-to-strong learning methods in human-designed sandboxes and evaluation testbeds. 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
Wen and colleagues with Claude: automated weak-to-strong supervision research
Ran parallel research agents on weak-to-strong supervision, obtaining and testing combinations of representation-based confidence filtering, evolutionary adapter refinement and posterior-label noise modeling, with selected method transfer to new math and coding datasets. Reported result; not independently verified by WRS. Ranked using a provisional grade and reference estimate for the exact claimed result.
Scientific scope
Repeated access makes the principal evaluation set effectively validation, and agents found shortcuts and evaluation exploits. Independent-domain transfer varies, and a production-scale posterior-label trial showed only a noise-floor improvement. Recovery on selected binary tasks is not a general solution to weak-to-strong alignment or superiority over unrestricted human research.
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
Jiaxin Wen; Liang Qiu; Joe Benton; Jan Hendrik Kirchner; Jan Leike
Named source authors and agent execution are recorded separately; individual effort shares are unknown.
Claude / Anthropic / Claude Opus 4.6: Source-reported autonomous research agents proposing, implementing and testing weak-to-strong learning methods in human-designed sandboxes and evaluation testbeds (source reported).