INDEX

Explanations

negative emotional experiences and reactions related to personal experiences

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

Following intensifiers (literally, almost, barely)

np_acts-logits-general · gemini-2.0-flash

literally zero

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 veldig

-0.77

djangoproject

-0.71

 appunto

-0.70

Very

-0.69

 <=",

-0.68

 dunque

-0.66

очень

-0.65

Sehr

-0.64

 väldigt

-0.63

 kuitenkin

-0.62

POSITIVE LOGITS

 practically

1.14

 literally

1.12

literally

1.07

 almost

1.07

 Literally

1.01

almost

0.98

Literally

0.98

 barely

0.98

 virtually

0.96

 nearly

0.94

Activations Density 0.463%

negative emotional experiences and reactions related to personal experiences

Following intensifiers (literally, almost, barely)

literally zero

No Comments

No Known Activations