INDEX

Explanations

phrases that introduce contrasting points or contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 21-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_21/width_16k/average_l0_70

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.21.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

olm

-0.55

 despite

-0.50

 obwohl

-0.48

despite

-0.45

 nonostante

-0.45

日まで

-0.44

ammen

-0.44

 sær

-0.43

 zanim

-0.43

one

-0.42

POSITIVE LOGITS

 Conversely

1.53

Conversely

1.41

 conversely

1.39

 Meanwhile

1.22

Meanwhile

1.19

 Whereas

1.16

 meanwhile

1.16

 Sedangkan

1.15

 natomiast

1.11

Likewise

1.10

Activations Density 0.368%

phrases that introduce contrasting points or contexts

No Comments

No Known Activations