baseball and sports mentions
A neuronOne of thousands of tiny dials inside each layer. It reads the text so far and produces a number, its activation, that nudges what the model writes next. in layerOne step in the stack the text passes through. Qwen3-0.6B has 28 layers, each with 3,072 neurons. 23 of Qwen3-0.6B. Its strongest snippets are mixed: they peak on different tokens. It is featuredA neuron picked for the full treatment: steering outputs and a written explanation. About 3,000 of the 86,000 neurons are featured., so it has steering outputs.
What this neuron does
This neuron lights up on baseball words, especially words like 'baseball', 'first base' and the words right before them. It mostly stays quiet or goes the other way on basketball and other sports talk. Turning it up did not make the model write about sports; the outputs were unrelated rambling.
Written by Claude from this neuron's activations and steering outputs. Confidence: medium. Think it is wrong? Flag it (opens a prefilled issue on GitHub).
Fires most on
The text windows where this neuron's activationThe number a neuron produces for a token. Large positive means it fired hard; negative means it pushed the other way. was highest. Amber marks positive activation; the outlined tokenA piece of text the model reads at once: often a word, sometimes part of a word or a punctuation mark. is the peak. Hover a token for its strength.
-
can bring a good deal of scrutiny and pressure. Under this pressure, Knoblauch began having problems with his throw to first base. Balls would sail over the head of the first baseman. He had thrown to first base thousands of times in
-
, linking Manhattan and Brooklyn, was opened.⏎In 1935 the major leagues' first night baseball game was played in Cincinnati. (The Reds defeated the Philadelphia Phillies 2-1.)⏎In 1941, in World
-
season with the Brooklyn Dodgers, most baseball fans and writers believed that he was the first black to play in the major leagues. (Robinson himself believed that at the time.) He was the fourth. William Edward White was
-
class who like baseball.⏎- I point to first base and ask them what it is. They will probably say "ファースト."⏎- I then ask him/her to say "second" and<|endoftext|>Swimming
-
, Shorty’s father decides to lead other Japanese Americans in building a baseball field at the camp and forming baseball teams. Shorty, who wasn’t a good baseball player back home, reaches a critical moment and must rise to the occasion. He
-
emy laws (Salman Taseer, the former governor of Punjab, and Shahbaz Bhatti, the Federal Minister
SteeringAdding a push along a neuron while the model writes, to see how its output changes. Turning it up or down.
Push this neuron up or down while the model writes, and watch what changes. The outputs were generated ahead of time, so nothing runs in your browser.
Here is something I have been thinking about lately: The question is, what is the point of a question? The question is a question. The question is a question. The question is a question. And the question is a question. The question is a question. What is the point of
Most negative on
Where the activation went furthest below zero. Blue marks negative activation. Many neurons mean something different in each direction.
-
not the last player this has happened to. Tennis players ‘forget’ how to toss the ball up in the serve. Basketball players clank free throws off the back of the backboard. Quarterbacks throw passes that sail off into the air.
-
of 245,811 points scored by 450 players (data here). The top 20% of scorers (led by #1 Kobe Bryant with 2,323 points) had 1
-
great deal to do with the success I have enjoyed as a man,” Peterson said in a statement.⏎Among those coming to Peterson’s defense was NBA Hall of Famer Charles Barkley. On the CBS pregame show The NFL Today, Barkley
-
... Olympic Basketball Tournament.⏎(Basketball term) the successful deflection of a shot by touching part of the ball on its way to the basket, thereby preventing a field goal⏎Một loại quần short ống rộng vừa vặn,
Connections
Show on mapWhich neuronsOne of thousands of tiny dials inside each layer. It reads the text so far and produces a number, its activation, that nudges what the model writes next. in earlier layersOne step in the stack the text passes through. Qwen3-0.6B has 28 layers, each with 3,072 neurons. feed this one, and which later ones it feeds. Amber links excite, blue links suppress; longer bars are stronger.
Fed by earlier layers
- L20:2469 excites
- L21:1039 suppresses
- L8:2451 excites
- L3:1552 excites
- L3:877 suppresses
- L10:735 suppresses
- L22:2216 suppresses
- L0:2525 excites
- L7:2260 excites
- L22:531 weakly suppresses sports leagues and team names
- L9:857 weakly excites
- L6:2384 weakly excites
- L6:1580 weakly suppresses
- L2:70 weakly excites
- L11:2987 weakly suppresses
- L1:1907 weakly suppresses
- L14:2747 weakly suppresses
- L3:2463 weakly suppresses
- L8:1573 weakly suppresses
- L14:941 weakly excites
- L9:332 weakly excites
- L9:845 weakly excites
- L1:247 weakly excites
- L3:1202 weakly excites
Feeds later layers
- L27:564 weakly suppresses
- L26:1855 weakly suppresses
- L24:147 weakly suppresses
- L24:1371 weakly excites
- L26:2467 weakly suppresses
- L24:1541 weakly suppresses
- L24:2606 weakly suppresses
- L25:2389 weakly excites
- L25:1576 weakly suppresses
- L27:1671 weakly suppresses
- L25:2276 weakly suppresses
- L24:1617 weakly suppresses
- L26:1273 weakly suppresses
- L24:1325 weakly excites
- L24:154 weakly suppresses
- L26:2428 weakly suppresses
- L24:1961 weakly excites
- L24:1664 weakly excites
- L24:284 weakly suppresses
- L25:1303 weakly excites
- L24:494 weakly excites
- L27:2723 weakly suppresses
- L24:1581 weakly excites
- L24:884 weakly excites
Activation histogramA chart of how often a neuron takes each activation value.
How often this neuron fires at each strength. Almost always it does close to nothing (the tall bars in the middle); the snippets above are the rare moments it fired hard, amber for positive and blue for negative.
Logit lensReading which output tokens a neuron pushes toward or away from, straight off its output weights, before the later layers have their say.
The words this neuron nudges the model toward (amber) or away from (blue) when it fires. A hint, not a label: many neurons work through later layersOne step in the stack the text passes through. Qwen3-0.6B has 28 layers, each with 3,072 neurons. instead.
Promotes
- itches
- 阪
- batter
- object
- b
- Objects
- objc
- açısından
- çi
- Chili
Suppresses
- {}",
- {}",
- guards
- 明珠
- guard
- ucks
- -guard
- []).
- cour
- kü