INDEX

Explanations

any

np_max-act · gemini-2.0-flash

The neuron lights up on semantically “heavy” content words (nouns, adjectives, etc.) rather than on function words.

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

')}

-0.07

ometimes

-0.07

.Web

-0.07

 Manip

-0.07

 EXCEPTION

-0.07

 servlet

-0.07

�

-0.06

 appe

-0.06

ipl

-0.06

 proprietary

-0.06

POSITIVE LOGITS

 Rakou

0.07

 Macedonia

0.06

�

0.06

IOD

0.06

 vazgeç

0.06

 Європ

0.06

voir

0.06

 zdję

0.06

 mixed

0.05

 дека

0.05

Activations Density 0.391%

any

The neuron lights up on semantically “heavy” content words (nouns, adjectives, etc.) rather than on function words.

No Comments

No Known Activations