- Rating
- 2
- Confidence
- 3
- Informativeness
- 2
- Presentation
- 3
- Soundness
- 2
Read review
Paper summary
This paper introduces CoDePass to identify role-related components by contrasting their contributions to complete reference responses under matched prompts. It analyzes attention-head and MLP patterns and reports gains on held-out RoleBench characters from fine-tuning selected attention components. The results support the usefulness of the ranking, but the proposed division of labor and improvements in character fidelity are not yet convincingly established.
Critical strengths
The selective fine-tuning results have meaningful controls. Ranking and training use 75 characters, while evaluation uses six disjoint characters. Random SFT matches layerwise head counts, shared GQA K/V groups, and parameter count, and the advantage persists across multiple training seeds and random selections. These comparisons provide useful evidence that the ranking identifies parameters whose adaptation transfers across characters.
Critical weaknesses
-
The proposed cue-to-identity-to-answer mechanism is stronger than the evidence. Figure 6a measures the pre-MLP residual, which combines attention and earlier MLP contributions, so this plot alone does not establish that MLPs produce the observed identity alignment. Figure 6b isolates MLP updates but evaluates their alignment with the response direction. Knockout and functional-class fine-tuning support component importance and tuning utility, but do not establish the claimed roles in identity selection, profile retrieval, and response timing.
-
The role-playing evaluation does not establish improved character fidelity. Selected cases in Figures 40–42 show Targeted SFT receiving higher ROUGE-L scores despite incorrect character motives or facts. The length controls and qualitative analysis are useful, but do not determine whether character knowledge and belief consistency improve across the held-out evaluation. This limits the use of the fine-tuning results as validation of the proposed role-playing mechanism.
Questions for authors
-
What evidence attributes the mid-layer increase in identity alignment to MLPs, given that Figure 6a probes the pre-MLP residual? Can the respective contributions of attention and MLP updates be separated, and can a targeted intervention establish a selective effect on role identity?
-
In Figure 41, Targeted SFT replaces Ripley’s fraudulent income with inheritance and investments yet obtains 33.33 ROUGE-L; Random Set 2 describes manipulation but scores 21.28. On the existing held-out outputs, does an assessment of character facts and beliefs, independent of reference overlap, still favor Targeted SFT?
-
Section 4.3 attributes Base-to-Instruct changes to instruction tuning, while Appendix C.5.2 also changes the completion interface to a chat interface. Do the conclusions persist under a shared template? What evidence establishes MLP stability beyond the nonsignificant test in Table 1?
If the authors can substantiate the proposed role-specific functions and confirm gains in character fidelity under an assessment independent of reference overlap, this could strengthen my confidence in the paper’s conclusions.
6 comments
Response to vTsp (2/3): W2; Q2
W2 & Q2. The role-playing evaluation does not establish improved character fidelity.
W2-1 & Q2: Character knowledge and belief consistency improve across the held-out evaluation.
An evaluation of character facts, motives, and beliefs on the existing held-out set—completely independent of reference overlap and combined with blinded human preference judgments—confirms that Targeted SFT significantly outperforms Random SFT and Full SFT.
First, reference-based lexical overlap (ROUGE-L) on RoleBench is a mainstream evaluation paradigm in recent role-playing literature, where factual content recovery and surface rhetorical style represent two distinct objectives. Recent studies—including ThinkPersona [1] and RoleCDE [2]—evaluate RoleBench using ROUGE-L.
Second, Figures 40–43 in Appendix E.6 were curated specifically from failure cases to transparently illustrate the shared conflict faced by safety-aligned LLMs when handling morally transgressive personas [2,4]. In fact, in Figure 41, Full SFT (25.64 R-L) and Random Sets 1 & 3 (30.30 R-L) commit the exact same error as Targeted SFT (replacing Ripley's fraudulent income with "inheritance and investments").
Finally, to test whether character knowledge and belief consistency improve without lexical-overlap bias, we re-evaluate the existing held-out outputs using an independent LLM judge (deepseek-flash, given only the character profile and question, without the reference answer) and blinded human evaluators (300 items, 3 annotators, ). Targeted SFT achieves the highest reference-free factual accuracy, belief/motive consistency, Overall score, and blinded human win rate among all models, and leads across all fine-tuned baselines:
| Method | Ref-Free Fact Score (1–5) ↑ | Belief / Motive (1–5) ↑ | Voice (1–5) ↑ | Overall ↑ | Targeted Human W/T/L |
|---|---|---|---|---|---|
| Qwen3-8B-Instruct | 3.10 ± 0.00 | 3.24 ± 0.00 | 3.42 ± 0.00 | 3.25 ± 0.00 | 172 / 77 / 51 |
| Full SFT | 3.26 ± 0.04 | 3.41 ± 0.03 | 3.12 ± 0.03 | 3.26 ± 0.03 | 127 / 107 / 66 |
| Random SFT | 3.21 ± 0.05 | 3.36 ± 0.05 | 3.15 ± 0.04 | 3.24 ± 0.04 | 184 / 56 / 60 |
| Targeted SFT | 3.54 ± 0.03 | 3.68 ± 0.02 | 3.29 ± 0.03 | 3.50 ± 0.02 | - |
Table C: Reference-free independent judge evaluation (1–5 scale) and blinded human preference.
W2-2: This limits the use of the fine-tuning results as validation of the proposed role-playing mechanism.
Our mechanistic conclusions rest on a three-part interpretability framework rather than fine-tuning alone, and the reference-free fidelity gains above further confirm that Targeted SFT validates the localized mechanism.
First, as recognized by Reviewers VRNr and ueGy, our primary contribution is the first systematic component-level mechanistic study of role playing in decoder-only LLMs—integrating CoDePass counterfactual attribution, progressive knockouts, and sublayer behavioral diagnostics.
Second, the proposed role-playing mechanism is established primarily through RQ1 (counterfactual localization and knockout necessity) and RQ2 (value-weighted interaction maps, isolated writes, and functional double dissociation in Tables A–B), rather than relying solely on RQ3 fine-tuning. Selective fine-tuning in RQ3 serves as a downstream prescriptive validation: combined with the reference-free fact and belief improvements in Table C, it indicates that the sparse attention components identified by CoDePass indeed govern character adaptation more effectively and efficiently (1.52% parameters, 35.9% faster training) than random or full-model updates.
Refs. [1] Cai: ThinkPersona (ACL'26); [2] Lai: RoleCDE (ACL'26); [3] Yan: Multi-component Tracing (ACL'26); [4] Yi: Too Good to be Bad (ACL'26); [5] Liao: MOA (ACL'26).
Response to vTsp (3/3): Q3
Q3. Section 4.3 attributes Base-to-Instruct changes to instruction tuning, while Appendix C.5.2 also changes the completion interface to a chat interface. Do the conclusions persist under a shared template? What evidence establishes MLP stability beyond the nonsignificant test in Table 1?
Q3-1: Shared prompt template between Base and Instruct.
The Base-to-Instruct reorganization conclusion holds fully when both checkpoints are evaluated under the exact same plain-text template (Shared Plain) or chat template (Shared Chat).
To rule out interface formatting as a confounder, we re-run the Jack Sparrow Base-vs.-Instruct attribution comparison (corresponding to Appendix Table 19, 100 identical questions per template) using identical input and target token sequences on both checkpoints. Table D confirms that under either shared template, attention heads undergo a highly significant distributional shift and top-set turnover, while MLP blocks maintain high overlap:
| Prompt Interface | Attention K-S D (p) | Attention Top-32 Overlap | MLP K-S D (p) | MLP Top-8 Overlap |
|---|---|---|---|---|
| Native (Table 19) | .584 (2.0e-182) | 17 / 32 (53.1%) | .556 (1.9e-5) | 6 / 8 (75.0%) |
| Shared Plain | .214 (<1e-20) | 16 / 32 (50.0%) | .222 (.340) | 6 / 8 (75.0%) |
| Shared Chat | .228 (<1e-22) | 15 / 32 (46.9%) | .250 (.212) | 6 / 8 (75.0%) |
Table D: Base-vs.-Instruct attribution comparisons under identical prompt templates.
Q3-2: Evidence for MLP stability beyond the two-sample K-S p-value.
Four sample-size-invariant rank correlation, cosine similarity, distributional divergence, and functional trajectory metrics consistently confirm that the MLP backbone is substantially more conserved and stable than attention heads.
First, as noted in Appendix C.5.1, because the two component families have different cardinalities (1,152 attention heads vs. 36 MLP blocks), we do not rely on the K-S test -value alone.
Second, Table E reports four cardinality-independent stability metrics under the shared plain-text interface—layerwise Spearman rank correlation (), attribution vector cosine similarity, Jensen-Shannon divergence (JSD), and direct sublayer-write identity trajectory correlation ( vs. )—which mutually reinforce our RQ3 fine-tuning finding (that updating a small fraction of attention parameters suffices, whereas updating all MLPs degrades performance, Table 24):
| Component Family | Spearman ρ ↑ | Attribution Cosine ↑ | JSD ↓ | Identity Trajectory r ↑ |
|---|---|---|---|---|
| Attention Heads | 0.3884 | 0.7267 | 0.1092 | 0.8332 |
| MLP Blocks | 0.7842 | 0.8914 | 0.0418 | 0.8578 |
Table E: Cardinality-invariant Base/Instruct stability metrics under the shared plain-text interface.
Refs. [1] Cai: ThinkPersona (ACL'26); [2] Lai: RoleCDE (ACL'26); [3] Yan: Multi-component Tracing (ACL'26); [4] Yi: Too Good to be Bad (ACL'26); [5] Liao: MOA (ACL'26).
Response to vTsp (1/3): W1; Q1
Thank you for recognizing the clarity of our presentation and the rigorous controls in our selective fine-tuning experiments—specifically that our 75-to-6 disjoint character split and layerwise/GQA-matched Random SFT confirm cross-character parameter transfer.
W1-1 & Q1: Isolating direct MLP updates (ΔMLP) from attention updates (ΔAttn) in identity alignment.
By decomposing the residual stream into isolated sublayer writes, we directly separate the contributions of attention mechanisms and MLP updates. This directly resolves the concern: middle-layer MLPs—not attention heads—are the primary drivers of the mid-layer surge in identity alignment and identity consolidation.
While Figure 6a plotted the pre-MLP residual state entering each block, we now separately compute the direct sublayer write projections cos(ΔAttnl,uT) and cos(ΔMLPl,uT) across depth (
120matched prompts across six characters), where ΔAttnl=Attnoutl−Attninl and ΔMLPl=MLPoutl−MLPinl. Table A shows that attention updates dominate early indicator/context routing (L0–L10), whereas direct MLP writes ΔMLPl drive the significant mid-layer surge in identity alignment (L11–L24, p<.001) and late-layer response generation (L25–L35):Table A: Cosine alignment of isolated sublayer writes (ΔAttnl vs. ΔMLPl) across depth.
W1-2 & Q1: Targeted causal intervention on mid-layer identity and routing functions.
Targeted sublayer interventions establish a clean functional double dissociation across profile retrieval, identity synthesis, and response-boundary timing.
Motivated by recent multi-component causal tracing research showing that functional circuits operate as coordinated component groups [3], we perform targeted interventions on each functional group across
200matched questions and compare against layer-matched random head/direction controls. Each intervention selectively impairs its corresponding target function without disrupting non-target capabilities (Table B): zeroing the 3 role-context heads selectively lowers profile content recall; zeroing the 2 role-indicator heads or removing the projection of mid-layer MLP writes (L12–L22) onto uID selectively reduces the Identity Probe score and LLM Judge score (while keeping content recall and normal boundary completion intact); and zeroing the positional heads selectively disrupts normal response-boundary termination while preserving persona quality:Table B: Dissociation effects of functional sublayer interventions vs. layer-matched random controls.
Therefore, targeted interventions establish a selective causal effect on role identity, and together with isolated sublayer writes, directly substantiate the claimed division of labor across profile retrieval, identity selection/synthesis, and response timing.
Refs. [1] Cai: ThinkPersona (ACL'26); [2] Lai: RoleCDE (ACL'26); [3] Yan: Multi-component Tracing (ACL'26); [4] Yi: Too Good to be Bad (ACL'26); [5] Liao: MOA (ACL'26).