tim ganiev (@postimortem) header
tim ganiev (@postimortem) profile picture

patterning personal intelligence

residual stream
Joined April 2014
Photo by tim ganiev (@postimortem)Photo by tim ganiev (@postimortem)Photo by tim ganiev (@postimortem)Photo by tim ganiev (@postimortem)Photo by tim ganiev (@postimortem)Photo by tim ganiev (@postimortem)Photo by tim ganiev (@postimortem)Photo by tim ganiev (@postimortem)Photo by tim ganiev (@postimortem)Photo by tim ganiev (@postimortem)
tim ganiev retweeted
Dostoevsky was right. Nothing kills the human soul more than being in the wrong place surrounded by the wrong people
289
28,673
946
160,804
2,634,945
tim ganiev retweeted
New neurons research: introducing Neuron Communities. We discovered communities of neurons that work together inside LLMs. They’re surprisingly small, but powerful – just a few neurons work together to perform common behaviors, like arithmetic, ALL CAPS, or recalling a date. Work led by research fellow @_igorshilov:
21
106
23
882
45,646
tim ganiev retweeted
tonight we're letting the amen break speak
43
1,318
54
12,452
152,575
tim ganiev retweeted
Claude's Creed
7
11
1
327
5,641
tim ganiev retweeted
Replying to @_mario_neo_
Imagine amino acids have feelings.
1
6
171
tim ganiev retweeted
Can interpretability shape training? @TIME covers our latest work on mechanistic shaping and why our team is optimistic. Read the story online or in TIME's upcoming print edition. time.com/article/2026/11/05/…
6
18
1
114
6,537
tim ganiev retweeted
We’ve been hard at work trying to make models work with harnesses better. The best thing @yb2698 found: “harness adapters”. Small LoRAs trained with RL inside the actual agent harness they’ll be used with. Same base model. Different adapter for pi, opencode, hermes, etc. Why? Simply because the same model behaves very differently depending on the tools, prompts and control flow around it. Instead of asking one set of weights to be optimal everywhere, we can specialize a tiny part of the model for each environment. Early results are pretty encouraging: on held-out tasks, the adapters make fewer tool calls than the base model while keeping or improving accuracy. And because they’re LoRAs, the operational side is surprisingly cheap too. For one LFM2.5-24B-A2B experiment, we evaluated 9 checkpoints / 1,350 agent sessions on a single H200 in under an hour. The total: ~$5! Merging and serving every checkpoint separately would have cost around $90. There’s another important piece: RL agents will eventually try to hack the reward. So every tool call goes through `safeagents`, a deterministic rule-based guard that blocks things like editing tests, tampering with rewards or bypassing the sandbox before they execute. And the whole loop can start from real usage: your coding sessions -> verifiable environments -> RL -> a better adapter for your agent. The model doesn’t have to change completely. Sometimes it just needs to get better at the harness it actually lives in. huggingface.co/blog/harness-…
38
34
7
172
27,693
If your cocaine dealer shorted you and didn’t supply you with the full amount of product, contact us at 336-373-2222 to file an official complaint.
5,574
34,167
6,524
357,287
18,199,141
tim ganiev retweeted
always start with evals. work from first principles
33
44
10
795
33,577
tim ganiev retweeted
people are like “why do LLMs always think they’re in an eval lol” and then have bad dreams about forgetting to turn in a homework assignment 10 years ago
78
351
16
6,421
135,472
tim ganiev retweeted
an announcement from the pierre computer company
381
138
161
4,004
1,196,504
tim ganiev retweeted
You should apply to this :) Fun personal fact: I applied to OpenAI residency like 8 times.
Applications are now open for the Anthropic Residency, a paid, fixed-term research program for academics, technical founders, and staff-level engineers who want to make AI safe and beneficial for everyone. No ML background is required. Apply by November 22: alignment.anthropic.com/2026…
51
25
1
1,486
191,706
tim ganiev retweeted
we gave qwen emergent misalignment by training it on bad advice in two separately trained models, the same neurons lit up on its misaligned answers we looked them up in my neuron atlas, and one theme kept coming up: power 🧵
13
20
3
209
10,338
tim ganiev retweeted
Schmidhuber already made this chart in 1996
1
1
139
tim ganiev retweeted
Replying to @alexandr_wang
108
99
47
7,597
374,712
Dudes would rather learn mechanistic interpretability rather than trying to move on (me)
found a neuron in qwen3.5-4b that makes it talk like her
5
17
400
8,961
tim ganiev retweeted
Replying to @inxanity06
it was never about the parameters
2
3
1
158
8,919
tim ganiev retweeted
79
53
24
2,163
101,519
tim ganiev retweeted
Replying to @typesafeai
I NEED THIS
1
1
10
505