INDEX

Explanations

contractions that express negation.

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

statements related to financial matters and technical details.

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

contractions with "not", particularly "isn't" and "wouldn't".

oai_token-act-pair · claude-3-5-sonnet-20240620 Triggered by @bot

"t" followed by a contraction of "would not"

np_acts-logits-general · gemini-2.0-flash

wouldn't, wasn't

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.20.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

“

-0.64

‘

-0.60

”

-0.53

”

-0.52

اه

-0.47

van

-0.47

 Persons

-0.46

 persons

-0.45

on

-0.45

спе

-0.45

POSITIVE LOGITS

Rhestr

1.03

 ་་

0.92

expandindo

0.91

 Charlemagne

0.90

"]]

0.90

IndentedString

0.86

 transfieras

0.86

 juſt

0.84

ListItemIcon

0.84

ьаж

0.84

Activations Density 0.099%

contractions that express negation.

statements related to financial matters and technical details.

contractions with "not", particularly "isn't" and "wouldn't".

"t" followed by a contraction of "would not"

wouldn't, wasn't

No Comments

No Known Activations