INDEX

Explanations

words related to negation or denial

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

Followed by "no" then a word expressing time

np_acts-logits-general · gemini-2.0-flash

no followed by adverb

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 22-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_22/width_16k/average_l0_72

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.22.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

 насељу

-0.47

 wenn

-0.47

zuführen

-0.45

 keď

-0.45

 whereas

-0.43

йдет

-0.42

parseFrom

-0.42

মান

-0.42

řízen

-0.42

 {}".

-0.41

POSITIVE LOGITS

tvguidetime

0.99

matically

0.91

 propOrder

0.82

fully

0.82

 ivelany

0.78

DockStyle

0.75

ically

0.75

tually

0.73

ingly

0.73

 likely

0.72

Activations Density 0.277%

words related to negation or denial

Followed by "no" then a word expressing time

no followed by adverb

No Comments

No Known Activations