Chelsea Finn ✓

@chelseabfinn · Palo Alto, CA · joined 19 Jun 2014

Asst Prof of CS & EE @Stanford Co-founder of Physical Intelligence @physical_int PhD from @Berkeley_EECS, EECS BS from @MIT

101 376Followers

397Following

704Posts total

510.2KViews on collected posts

Postingan terbaru

Chelsea Finn ✓@chelseabfinn23 Aug

One of the most important aspects of scientific discovery is deciding where to draw insights from. While LLMs are promising tools for science, we lack datasets & evaluations for this step. Help contribute to a public dataset for exactly this: https://t.co/X8NPVnVIwm 51.7K views · 267 likes · 22 reposts · 7 replies Open on X →

Chelsea Finn ✓@chelseabfinn17 Aug

We’re asking the research community to help us build a benchmark for research taste. Scientific discovery starts with a fundamental step: which prior work is worth building on? We want to capture this undocumented layer through our collective knowledge. Please sign up: https:// 72.9K views · 208 likes · 36 reposts · 5 replies Open on X →

Chelsea Finn ✓@chelseabfinn9 Aug

@chelseabfinn makes sense, narrow-policy data can bake in a brittle value landscape before finetuning starts. 170 views · 1 likes · 0 reposts · 0 replies Open on X →

Chelsea Finn ✓@chelseabfinn9 Aug

@chelseabfinn so the Q-function overfits to the pretraining policy's distribution unless you mix in data from other policies how many distinct policies did you need in the dataset before it started helping? 738 views · 3 likes · 0 reposts · 0 replies Open on X →

Chelsea Finn ✓@chelseabfinn9 Aug

@chelseabfinn what policy diversity did you test? like random vs expert vs noisy 450 views · 1 likes · 0 reposts · 0 replies Open on X →

Chelsea Finn ✓@chelseabfinn9 Aug

Pretraining a Q-function often doesn’t actually help RL finetuning, compared to initializing Q from scratch. We find that pretraining Q-functions on data from diverse policies is critical to see improvements from pretraining. Paper: https://t.co/dlj5RXVFED 78K views · 641 likes · 54 reposts · 13 replies Open on X →

Chelsea Finn ✓@chelseabfinn30 Jul

Pretraining has worked remarkably well across domains We show this doesn’t hold for Q-functions in online RL from a pretrained policy — and propose IPE, a more effective way to learn Q-functions for online RL fine-tuning (1/6) https://t.co/G5SYHmgz83 83.4K views · 305 likes · 29 reposts · 5 replies Open on X →

Chelsea Finn ✓@chelseabfinn9 Jul

I'm giving a talk tomorrow at ICML on emergent physical generalization, including π0.7 🤖 3:15 pm @ SCALE workshop in Ballroom 201 https://t.co/xppCOkktOJ 27.9K views · 369 likes · 20 reposts · 12 replies Open on X →

Chelsea Finn ✓@chelseabfinn9 Jul

I'm giving a talk on how we can move beyond the scalar reward bottleneck for both robotics & LLMs. ICML RLxF workshop tomorrow at 1:30 pm. 53.2K views · 276 likes · 21 reposts · 9 replies Open on X →

Chelsea Finn ✓@chelseabfinn9 Jul

RL is hitting a ceiling with human feedback. What if the world itself becomes the signal? Join us at the RLxF: RL from World Feedback 🌍 workshop at ICML 2026 @icmlconf tomorrow (July 10th)! Web page: https://t.co/cN0itnL1yI https://t.co/vpMTQe4VCm 69.4K views · 68 likes · 11 reposts · 1 replies Open on X →

Chelsea Finn ✓@chelseabfinn16 Apr

LLM post-training used to mean fine-tuning to a downstream task Robotics has been stuck in this setting, needing task-specific fine-tuning for best performance π07 changes this: It works out of the box & outperforms fine-tuned specialists Details: https://t.co/QbO3E4D3QN https 72.5K views · 617 likes · 64 reposts · 29 replies Open on X →

Dibandingkan akun berukuran sama

6 postingan dari 90 hari terakhir, dibandingkan dengan rentang 100K–1M pengikut. tampil luas, tetapi sedikit penonton yang merespons.

Median tayangan26 216akun ini5 983median untuk 100K–1M

Jangkauan, %25.86%akun ini1.49%median untuk 100K–1M

Interaksi, %0.49%akun ini0.93%median untuk 100K–1M

Metrik	Akun ini	Median untuk 100K–1M	Rasio
Median tayangan per postingan	26 216	5 983	4.38×
Jangkauan (tayangan ÷ pengikut)	25.86%	1.49%	17.4×
Tingkat interaksi	0.49%	0.93%	0.53×

Akun lain pada rentang ini → Bandingkan dengan akun lain → Bagaimana tolok ukur ini disusun →

Growth & engagement

How the posts we collected actually performed: views and reaction rate post by post, what the audience did with them, and where the follower count goes.

Views per post

72.5K16 Apr

69.4K9 Jul

53.2K

27.9K

83.4K30 Jul

78K9 Aug

450

738

170

72.9K17 Aug

51.7K23 Aug

Last 11 collected posts, oldest on the left. The scale is logarithmic: one post can outrun the rest a hundred times over.

Engagement rate per post

1.00%16 Apr

0.12%9 Jul

0.57%

1.44%

0.41%30 Jul

0.92%9 Aug

0.22%

0.41%

0.59%

0.35%17 Aug

0.57%23 Aug

Reactions — likes, reposts, replies and quotes — divided by views. Median for 100K–1M accounts is 0.93%.

What the audience does

Likes59.6%2 756 in total

Reposts5.6%257 in total

Replies1.8%81 in total

Quotes0.7%34 in total

Bookmarks32.4%1 500 in total

Share of every reaction we collected for this account. Replies mean argument, reposts mean endorsement, bookmarks mean the post was worth keeping.

The follower curve appears once this account has two daily snapshots — we take one a day, and this one is on its first.

Postingan terbaru

Dibandingkan akun berukuran sama

Growth & engagement

Views per post

Engagement rate per post

What the audience does

Akun serupa