responsible investing and ESG terms
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. 6 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 responds to business and money words about doing good, such as the 'Responsible' in Socially Responsible Investing or 'chain' in supply chain. When turned up, the model tended to drift toward talk about companies, climate and sustainability.
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.
-
didn’t realize it at the time, but he was already practicing socially responsible investing.⏎What is Socially Responsible Investing?⏎Socially responsible investing (sometimes also called Ethical Investing) is exactly
-
will also skip companies who don’t engage in socially responsible behavior (like coffee/restaurant companies who don’t use fair trade beans, for example). Socially responsible investing (SRI) is basically the practice of putting your money to work only in
-
+ community. As we’ll discuss in a minute, you are free to make your own choices about what is (and isn’t) socially responsible.⏎Ethical Investing is More Popular Than Ever⏎These days, there are a growing number of people
-
that can record transactions and information in a secure and transparent manner.⏎Blockchain has the potential to disrupt industries that rely on trust and verification, such as finance, supply chain management, and voting systems. It eliminates the need for intermediaries and enables secure
-
curbs on emissions, for example an international carbon price. A carbon price of $50 per tonne emitted is already accounted for by some companies, CDP said.⏎If it were to be introduced
-
peace Canada delegation arrived in Oslo Monday. Two of the delegates, Andrew Nikiforuk and Dr. John O'Connor, have raised concerns about the environmental, social and financial risks of tar sands investments that are reverberating through the Nordic business
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: 20.5, 20.5, 20.5, 20.5, 20.5, 20.5, 20.5, 20.5,
Most negative on
Where the activation went furthest below zero. Blue marks negative activation. Many neurons mean something different in each direction.
-
U.N) shows that, while prices are once again down from a peak, a troublesome volatility started to affect food prices this decade. These are the very
-
form and as a result it begins to crystallize. It doesn’t affect the performance of the candle in any way and also means your wax<|endoftext|>Equalization - Sound Mixing.⏎May
-
are betting that the price of the security will rise, allowing you to get a good deal later. By selling futures contracts, you are indicating that you are bearish about Bitcoin and that you predict a price decline.⏎- Prediction Markets⏎Shorting
-
pickup in “headline” inflation. However, core inflation has also increased; the year-over-year core CPI (excluding volatile food and energy prices) moved from a record low of 0.6 percent in October 2010 to
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
- L3:1346 suppresses
- L4:2409 excites
- L4:2651 suppresses
- L4:2954 suppresses
- L5:91 weakly excites
- L3:2584 weakly suppresses
- L3:928 weakly excites
- L5:258 weakly excites
- L3:464 weakly excites
- L2:2261 weakly suppresses
- L4:1671 weakly excites
- L3:258 weakly suppresses
- L5:50 weakly excites
- L5:353 weakly suppresses
- L1:2804 weakly excites
- L3:1702 weakly suppresses
- L4:2414 weakly excites
- L5:748 weakly excites
- L2:1494 weakly suppresses
- L5:96 weakly suppresses
- L3:66 weakly excites
- L3:1337 weakly suppresses
- L5:2135 weakly excites
- L0:4 weakly suppresses
Feeds later layers
- L7:2399 suppresses
- L7:1999 excites
- L8:2769 suppresses
- L7:1857 excites
- L10:1683 excites
- L7:1291 excites
- L7:1259 excites
- L7:2603 suppresses
- L7:2108 suppresses
- L15:2970 excites
- L19:2752 excites
- L10:2048 weakly suppresses
- L7:1303 weakly suppresses
- L7:1030 weakly suppresses
- L9:2662 weakly excites
- L11:1869 weakly excites
- L18:2298 weakly suppresses
- L9:1548 weakly suppresses
- L9:1018 weakly suppresses
- L22:1444 weakly suppresses
- L8:2711 weakly suppresses
- L20:1864 weakly suppresses
- L9:772 weakly suppresses
- L7:2930 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
- uts
- _IRQn
- uri
- source
- .jsx
- yard
- stry
- <source
- .names
- upro
Suppresses
- ставлен
- pressions
- SETTING
- Kou
- DropDown
- ировки
- тем
- RESET
- ifferential
- olley