INDEX

Explanations

Colors and shades

np_max-act · gemini-2.0-flash

This neuron responds to occurrences of the Russian word “тональность” (i.e. references to the tone or sentiment being positive/negative).

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

mik

-0.07

 dismant

-0.06

 parsers

-0.06

 uděl

-0.06

 کند

-0.06

xaa

-0.06

Як

-0.06

。しかし

-0.06

 Grab

-0.06

 mower

-0.06

POSITIVE LOGITS

 Eleven

0.08

 coupon

0.07

 Panasonic

0.06

Le

0.06

 Northern

0.06

 association

0.06

(phase

0.06

 align

0.06

 niż

0.06

Publish

0.06

Activations Density 0.000%

Colors and shades

This neuron responds to occurrences of the Russian word “тональность” (i.e. references to the tone or sentiment being positive/negative).

No Comments

No Known Activations

Colors and shades

This neuron responds to occurrences of the Russian word “тональность” (i.e. references to the tone or sentiment being positive/negative).

No Comments

No Known Activations