INDEX

Explanations

ensive/intensive

np_max-act · gemini-2.0-flash

terms related to intensive use and treatment methods, particularly in healthcare and therapeutic contexts.

oai_token-act-pair · gpt-4o-mini Triggered by @xinyanhu8

The neuron specifically activates on occurrences of the word “intensive.”

oai_token-act-pair · o4-mini Triggered by @xinyanhu8

New Auto-Interp

Configuration

andyrdt/saes-llama-3.1-8b-instruct/resid_post_layer_11/trainer_1

Dataset (Dashboard)

Various

Features

131,072

Data Type

float32

Hook Name

blocks.11.hook_resid_post

Architecture

standard

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 emits

-0.06

 wholly

-0.06

 відбувається

-0.06

"]=="

-0.06

 kurtar

-0.06

.↵↵↵↵↵↵↵↵↵↵

-0.06

일

-0.06

 mostr

-0.06

宽

-0.06

/hooks

-0.06

POSITIVE LOGITS

 intensive

0.11

ennie

0.08

-intensive

0.08

Thi

0.08

 Processing

0.07

 intensified

0.07

по

0.07

sis

0.07

nie

0.07

итор

0.07

Activations Density 0.002%

ensive/intensive

terms related to intensive use and treatment methods, particularly in healthcare and therapeutic contexts.

The neuron specifically activates on occurrences of the word “intensive.”

No Comments

No Known Activations