INDEX

Explanations

phrases and clauses involving contrasting ideas or situations

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

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

ichè

-0.47

 whereby

-0.46

choon

-0.46

estors

-0.46

downe

-0.45

 INTERESAR

-0.45

wäh

-0.43

หว่าง

-0.43

逅

-0.43

 SafeMath

-0.43

POSITIVE LOGITS

sort

1.03

 Literally

1.03

 sort

0.99

 literally

0.98

literally

0.98

Literally

0.98

 literalmente

0.97

 really

0.92

really

0.92

Sort

0.84

Activations Density 0.228%

phrases and clauses involving contrasting ideas or situations

No Comments

No Known Activations