INDEX

Explanations

treatments that work or don't work

np_acts-logits-general · gemini-2.5-flash-lite

words and phrases related to medications and their effects.

oai_token-act-pair · claude-3-7-sonnet-20250219 Triggered by @neilrathi

This neuron responds strongly to common, non‐specific function words and verbs (e.g. “will,” “it,” “do,” “can,” “stuff”) that tend to appear in vague or filler contexts.

oai_token-act-pair · o4-mini Triggered by @jyhe0408

New Auto-Interp

Configuration

google/gemma-scope-27b-pt-res/layer_34/width_131k

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 jambes

-0.99

lies

-0.96

 suffer

-0.96

>−</

-0.95

receive

-0.94

 позволяет

-0.86

itness

-0.85

şı

-0.84

uze

-0.83

ainte

-0.83

POSITIVE LOGITS

 cause

1.43

 help

1.37

 contains

1.31

 helped

1.20

 contain

1.19

 give

1.16

 provide

1.13

 causes

1.11

 caused

1.11

 causing

1.07

Activations Density 0.052%

treatments that work or don't work

words and phrases related to medications and their effects.

This neuron responds strongly to common, non‐specific function words and verbs (e.g. “will,” “it,” “do,” “can,” “stuff”) that tend to appear in vague or filler contexts.

No Comments

No Known Activations

treatments that work or don't work

words and phrases related to medications and their effects.

This neuron responds strongly to common, non‐specific function words and verbs (e.g. “will,” “it,” “do,” “can,” “stuff”) that tend to appear in vague or filler contexts.

No Comments

No Known Activations