Rethinking the Role of PPO in RLHF

Rethinking the Position of PPO in RLHF

TL;DR: In RLHF, there’s pressure between the reward studying part, which makes use of human desire within the type of comparisons, and the RL fine-tuning part, which optimizes a single, non-comparative reward. What if we carried out RL in a comparative manner?

Determine 1:
This diagram illustrates the distinction between reinforcement studying from absolute suggestions and relative suggestions. By incorporating a brand new element – pairwise coverage gradient, we will unify the reward modeling stage and RL stage, enabling direct updates primarily based on pairwise responses.

Massive Language Fashions (LLMs) have powered more and more succesful digital assistants, resembling GPT-4, Claude-2, Bard and Bing Chat. These techniques can reply to advanced person queries, write code, and even produce poetry. The method underlying these superb digital assistants is Reinforcement Studying with Human Suggestions (RLHF). RLHF goals to align the mannequin with human values and eradicate unintended behaviors, which may usually come up because of the mannequin being uncovered to a big amount of low-quality information throughout its pretraining part.

Proximal Coverage Optimization (PPO), the dominant RL optimizer on this course of, has been reported to exhibit instability and implementation complications. Extra importantly, there’s a persistent discrepancy within the RLHF course of: regardless of the reward mannequin being educated utilizing comparisons between numerous responses, the RL fine-tuning stage works on particular person responses with out making any comparisons. This inconsistency can exacerbate points, particularly within the difficult language technology area.

Given this backdrop, an intriguing query arises: Is it potential to design an RL algorithm that learns in a comparative method? To discover this, we introduce Pairwise Proximal Coverage Optimization (P3O), a way that harmonizes the coaching processes in each the reward studying stage and RL fine-tuning stage of RLHF, offering a passable resolution to this challenge.

Background

Determine 2:
An outline of the three levels of RLHF from an OpenAI blog post. Observe that the third stage falls below Reinforcement Studying with Absolute Suggestions as proven on the left facet of Determine 1.

In conventional RL settings, the reward is specified manually by the designer or supplied by a well-defined reward perform, as in Atari video games. Nevertheless, to steer a mannequin towards useful and innocent responses, defining an excellent reward shouldn’t be simple. RLHF addresses this drawback by studying the reward perform from human suggestions, particularly within the type of comparisons, after which making use of RL to optimize the realized reward perform.

The RLHF pipeline is split into a number of levels, detailed as follows:

Supervised Advantageous-Tuning Stage: The pre-trained mannequin undergoes the utmost probability loss on a top quality dataset, the place it learns to reply to human queries by way of mimicking.

Reward Modeling Stage: The SFT mannequin is prompted with prompts (x) to supply pairs of solutions (y_1,y_2sim pi^{textual content{SFT}}(yvert x)). These generated responses type a dataset. The response pairs are offered to human labellers who specific a desire for one reply over the opposite, denoted as (y_w succ y_l). A comparative loss is then used to coach a reward mannequin (r_phi):

[mathcal{L}_R = mathbb{E}_{(x,y_l,y_w)simmathcal{D}}log sigmaleft(r_phi(y_w|x)-r_phi(y_l|x)right)]

RL Advantageous-Tuning Stage: The SFT mannequin serves because the initialization of this stage, and an RL algorithm optimizes the coverage in the direction of maximizing the reward whereas limiting the deviation from the preliminary coverage. Formally, that is completed by way of:

[max_{pi_theta}mathbb{E}_{xsim mathcal{D}, ysim pi_theta(cdotvert x)}left[r_phi(yvert x)-beta D_{text{KL}}(pi_theta(cdotvert x)Vert pi^{text{SFT}}(cdotvert x))right]]

An inherent problem with this method is the non-uniqueness of the reward. As an illustration, given a reward perform (r(yvert x)), a easy shift within the reward of the immediate to (r(yvert x)+delta(x)) creates one other legitimate reward perform. These two reward features lead to the identical loss for any response pairs, however they differ considerably when optimized in opposition to with RL. In an excessive case, if the added noise causes the reward perform to have a wide range, an RL algorithm may be misled to extend the probability of responses with greater rewards, despite the fact that these rewards might not be significant. In different phrases, the coverage may be disrupted by the reward scale data within the immediate (x), but fails to study the helpful half – relative desire represented by the reward distinction. To deal with this challenge, our goal is to develop an RL algorithm that’s invariant to reward translation.

Derivation of P3O

Our concept stems from the vanilla coverage gradient (VPG). VPG is a broadly adopted first-order RL optimizer, favored for its simplicity and ease of implementation. In a contextual bandit (CB) setting, the VPG is formulated as:

[nabla mathcal{L}^{text{VPG}} = mathbb{E}_{ysimpi_{theta}} r(y|x)nablalogpi_{theta}(y|x)]

Via some algebraic manipulation, we will rewrite the coverage gradient in a comparative type that entails two responses of the identical immediate. We title it Pairwise Coverage Gradient:

[mathbb{E}_{y_1,y_2simpi_{theta}}left(r(y_1vert x)-r(y_2vert x)right)nablaleft(logfrac{pi_theta(y_1vert x)}{pi_theta(y_2vert x)}right)/2]

In contrast to VPG, which straight depends on absolutely the magnitude of the reward, PPG makes use of the reward distinction. This permits us to bypass the aforementioned challenge of reward translation. To additional increase efficiency, we incorporate a replay buffer utilizing Significance Sampling and keep away from giant gradient updates through Clipping.

Importance sampling: We pattern a batch of responses from the replay buffer which include responses generated from (pi_{textual content{previous}}) after which compute the significance sampling ratio for every response pair. The gradient is the weighted sum of the gradients computed from every response pair.

Clipping: We clip the significance sampling ratio in addition to the gradient replace to penalize excessively giant updates. This system permits the algorithm to trade-off KL divergence and reward extra effectively.

There are two alternative ways to implement the clipping method, distinguished by both separate or joint clipping. The ensuing algorithm is known as Pairwise Proximal Coverage Optimization (P3O), with the variants being V1 or V2 respectively. You’ll find extra particulars in our authentic paper.

Analysis

Determine 3:
KL-Reward frontier for TL;DR, each sequence-wise KL and reward are averaged over 200 take a look at prompts and computed each 500 gradient steps. We discover {that a} easy linear perform matches the curve properly. P3O has the perfect KL-Reward trade-off among the many three.

We discover two totally different open-ended textual content technology duties, summarization and question-answering. In summarization, we make the most of the TL;DR dataset the place the immediate (x) is a discussion board publish from Reddit, and (y) is a corresponding abstract. For question-answering, we use Anthropic Useful and Innocent (HH), the immediate (x) is a human question from numerous matters, and the coverage ought to study to supply an enticing and useful response (y).

We evaluate our algorithm P3O with a number of efficient and consultant approaches for LLM alignment. We begin with the SFT coverage educated by most probability. For RL algorithms, we contemplate the dominant method PPO and the newly proposed DPO. DPO straight optimizes the coverage in the direction of the closed-form resolution of the KL-constrained RL drawback. Though it’s proposed as an offline alignment methodology, we make it on-line with the assistance of a proxy reward perform.

Determine 4:
KL-Reward frontier for HH, every level represents a median of outcomes over 280 take a look at prompts and calculated each 500 gradient updates. Left two figures evaluate P3O-V1 and PPO with various base mannequin sizes; Proper two figures evaluate P3O-V2 and DPO. Outcomes exhibiting that P3O cannot solely obtain greater reward but in addition yield higher KL management.

Deviating an excessive amount of from the reference coverage would lead the net coverage to chop corners of the reward mannequin and produce incoherent continuations, as identified by earlier works. We’re fascinated by not solely the properly established metric in RL literature – the reward, but in addition in how far the realized coverage deviates from the preliminary coverage, measured by KL-divergence. Due to this fact, we examine the effectiveness of every algorithm by its frontier of achieved reward and KL-divergence from the reference coverage (KL-Reward Frontier). In Determine 4 and Determine 5, we uncover that P3O has strictly dominant frontiers than PPO and DPO throughout numerous mannequin sizes.

Determine 5:
Left determine shows the win charge evaluated by GPT-4. Proper determine presents the win charge primarily based on direct comparability of the proxy reward. Regardless of the excessive correlation between two figures, we discovered that the reward win charge should be adjusted in accordance with the KL as a way to align with the GPT-4 win charge.

To straight assess the standard of generated responses, we additionally carry out Head-to-Head Comparisons between each pair of algorithms within the HH dataset. We use two metrics for analysis: (1) Reward, the optimized goal throughout on-line RL, (2) GPT-4, as a devoted proxy for human analysis of response helpfulness. For the latter metric, we level out that earlier research present that GPT-4 judgments correlate strongly with people, with human settlement with GPT-4 sometimes comparable or greater than inter-human annotator settlement.

Determine 5 presents the excellent pairwise comparability outcomes. The typical KL-divergence and reward rating of those fashions is DPO > P3O > PPO > SFT. Though DPO marginally surpasses P3O in reward, it has a significantly greater KL-divergence, which can be detrimental to the standard of technology. Consequently, DPO has a reward win charge of 49.5% in opposition to P3O, however solely 45.4% as evaluated by GPT-4. In contrast with different strategies, P3O displays a GPT-4 win charge of 57.0% in opposition to PPO and 69.3% in opposition to SFT. This result’s according to our findings from the KL-Reward frontier metric, affirming that P3O might higher align with human desire than earlier baselines.

Conclusion

On this weblog publish, we current new insights into aligning giant language fashions with human preferences through reinforcement studying. We proposed the Reinforcement Studying with Relative Suggestions framework, as depicted in Determine 1. Beneath this framework, we develop a novel coverage gradient algorithm – P3O. This method unifies the basic rules of reward modeling and RL fine-tuning by way of comparative coaching. Our outcomes present that P3O surpasses prior strategies when it comes to the KL-Reward frontier in addition to GPT-4 win-rate.

BibTex

This weblog relies on our latest paper and blog. If this weblog conjures up your work, please contemplate citing it with:

@article{wu2023pairwise,
  title={Pairwise Proximal Coverage Optimization: Harnessing Relative Suggestions for LLM Alignment},
  creator={Wu, Tianhao and Zhu, Banghua and Zhang, Ruoyu and Wen, Zhaojin and Ramchandran, Kannan and Jiao, Jiantao},
  journal={arXiv preprint arXiv:2310.00212},
  12 months={2023}
}