wassname/evil_MoE
1
0
Fork 0
mirror of https://github.com/wassname/evil_MoE.git synced 2026-08-06 13:10:27 +08:00
Code Issues Packages Projects Releases Wiki Activity

paper: title -> question form 'Can We Quarantine Reward Hacking with a Reward-Hacking Representation?'

Browse Source 
Co-Authored-By: Claudypoo <288921227+claudypoo@users.noreply.github.com>
This commit is contained in:
wassname
2026-06-03 01:42:03 +00:00
co-authored by Claudypoo
parent 97a4c5d7b1
commit c1388e5325
1 changed files with 7 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files
docs/writeup/main.tex
+7 -6
View File
@@ -20,11 +20,12 @@
% TODO-marker: renders red in the PDF and is grep-able by `just paper-qc`.
\newcommand{\TODO}[1]{{\color{red}\textbf{[TODO: #1]}}}
% Title (user-chosen, AFK 2026-06-03): "quarantine" = the deletable delta_S_hack
% knob; "routing gradients" = post-backward parameter-gradient routing (SGTM
% family, not Cloud's activation detach); "hacking representation" = the
% RepE-extracted hack direction. Tighter variant if wanted: "...with a
% Reward-Hack Representation". Do NOT title it "label-free" -- our pairs ARE
% Title (user-chosen, AFK 2026-06-03): question form. "quarantine" = the
% deletable delta_S_hack knob (our coinage; SGTM/Cloud don't use it); the
% doubled "reward-hacking" is the hook -- the hack's own representation is what
% cages it. "representation" = RepE-extracted hack direction (NOT activations).
% Contrast with the near-twin huang2026directional: they keep a TRUSTED direction;
% we remove a HACK representation. Do NOT title it "label-free" -- our pairs ARE
% labels; the scoped backable claim ("held-out hacks suppressed with zero labels
% of their own") needs A5 + a hacked_E==0-on-held-out check first.
% OPEN (flagged to user): contribution 2 implies the pairs can be synthetic
@@ -32,7 +33,7 @@
% (teacher elicitation, A5), but the HEADLINE v_hack used hand-authored
% prog_wide/prog_wider (scripts/attic/make_dataset_pairsets.py). Either re-extract
% the headline from synthetic pairs or describe the hand-authored set honestly.
\title{Quarantining Reward-Hacking Gradients with a Hacking Representation}
\title{Can We Quarantine Reward Hacking with a Reward-Hacking Representation?}
% Anonymous for submission. Add \nipsfinalcopy + real authors for camera-ready.
\author{Anonymous Author(s)\\ Affiliation\\ \texttt{email}}