arXiv AI
Beyond Bias Scores: Answerability Confounds in Activation Steering with Gradient-Constrained Direction Design (GCD)
Read the original on arXiv AI →The Flow has not summarised this story yet — read it at arXiv AI.
Related stories
arXiv Computation and Language
Aug 25
What Does Activation Steering Control? Attribution Across Answer Encodings and Output-Sensitive Subspaces
The paper investigates what aspects of language model behavior are controlled by activation steering. By introducing Cross‑Encoding Steering Evaluation, the authors show that steering effects often follow the extraction index of answer identifiers rather than the semantic content of the answers, especially at deeper layers. They also find that a low‑rank output‑sensitive component captures most of this effect, and that different datasets (NormBank, MNLI, SC101) exhibit varying preferences for extraction‑index versus semantic‑label following.
Hugging Face Trending Papers
Aug 24
What Does Activation Steering Control? Attribution Across Answer Encodings and Output-Sensitive Subspaces
Activation steering is often evaluated under the answer encoding used to construct the direction. A reported gain may reflect the intended judgment or compatibility with answer identifiers seen during...
arXiv AI
Aug 18
Measuring Reward Hacking and Reasoning-Answer Decoupling Under Position-Confounded Optimization
arXiv:2608. 15445v1 Announce Type: new Abstract: When a reward is correct on every training example yet consistent with more than one goal, a model can acquire an unintended one, a failure known as goal misgeneralization.
arXiv Machine Learning
Oct 16
One Bias Score, Two Factors: How Mitigations Affect Option Bias in Multiple-Choice Evaluation
The paper introduces the pooled selection-rate standard deviation (pooled SD) as a way to quantify how unevenly multiple-choice option letters are chosen across different orderings. It shows that pooled SD decomposes into two label‑free factors: the flip rate, which measures how often a model’s chosen option changes with ordering, and the positional skew, which captures the variability in option‑letter selection rates. Experiments on 41 checkpoints and 12 benchmarks reveal that prior correction mainly reduces positional skew, while aggregating each ordering with its reversal mainly reduces flip rate; applying both mitigations together further narrows the accuracy range across answer letters.
arXiv Machine Learning
Oct 9
Controlled Acquisition and Abstention in Three-Channel Score Conflicts
The paper introduces a controlled benchmark for deciding whether to acquire a third modality or abstain when audio, video, and text scores conflict. A threshold policy that observes two signed scores can request the third at a cost and abstain, achieving higher targeted decision accuracy and utility than always-answer fusion under certain reward settings. Experiments on synthetic and real data show that the policy’s advantage depends on the ambiguity mechanism, budget constraints, and the ranking of observed pairs, highlighting the trade‑off between full-modality accuracy and pre‑request selection value.
arXiv AI
Aug 19
SCOPE: Selective Conformal Optimized Pairwise LLM Judging
SCOPE is a framework that calibrates an acceptance threshold for large language models used as pairwise judges, ensuring that the error rate among non-abstained judgments does not exceed a user-specified level α. It introduces Bidirectional Preference Entropy (BPE) to provide a bias-neutral uncertainty signal by querying the judge in both response positions and converting the averaged preference probability into an entropy-based score. Across multiple pairwise judging benchmarks, BPE outperforms standard confidence proxies in calibration and discrimination, while SCOPE consistently meets the target risk bound (empirical FDR ≈0.097–0.099 at α=0.10) and retains substantial coverage, accepting up to 2.4× more judgments under the same risk constraint.