INDEX

Explanations

question answering

np_max-act · gemini-2.0-flash

The neuron activates on tokens that mark user questions (interrogative words like “is,” “who,” “what,” “where,” or their equivalents in other languages, plus the trailing question mark).

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

BPM

-0.07

vekili

-0.07

 dine

-0.06

.mdl

-0.06

až

-0.06

<meta

-0.06

]";↵

-0.06

-site

-0.06

istes

-0.06

致

-0.06

POSITIVE LOGITS

 irres

0.06

_secure

0.06

 زمینه

0.06

 лич

0.06

 CircularProgressIndicator

0.06

 realloc

0.06

Vk

0.06

 İngiliz

0.06

pee

0.06

Specify

0.06

Activations Density 0.026%

question answering

The neuron activates on tokens that mark user questions (interrogative words like “is,” “who,” “what,” “where,” or their equivalents in other languages, plus the trailing question mark).

No Comments

No Known Activations

question answering

The neuron activates on tokens that mark user questions (interrogative words like “is,” “who,” “what,” “where,” or their equivalents in other languages, plus the trailing question mark).

No Comments

No Known Activations