INDEX

Explanations

negative expressions about respect and opinions in social contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 5-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_5/width_16k/average_l0_68

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.5.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

mektedir

-0.80

maktadır

-0.64

rsiniz

-0.57

<eos>

-0.52

 almendras

-0.47

 venons

-0.47

awaiter

-0.47

 soggetto

-0.46

cination

-0.45

美味しかったです

-0.45

POSITIVE LOGITS

 Савезне

0.94

InputBorder

0.84

 itſelf

0.82

󠁿

0.81

новниш

0.80

ſelf

0.78

:]:

0.78

 faſt

0.77

)':

0.76

 transfieras

0.75

Activations Density 0.329%

negative expressions about respect and opinions in social contexts

No Comments

No Known Activations