INDEX

Explanations

elements related to communication and interaction in discussions

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 34-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_34/width_16k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.34.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

!*\

-0.42



-0.38

værende

-0.36

rungsseite

-0.36

ljiv

-0.35

おり

-0.33

ArrowToggle

-0.32

uç

-0.30

iocene

-0.30

dolu

-0.29

POSITIVE LOGITS

 then

4.59

 Then

3.39

Then

3.38

then

3.36

 entonces

3.20

 então

3.05

 THEN

3.05

 dann

3.00

 سپس

2.67

 kemudian

2.66

Activations Density 4.536%

elements related to communication and interaction in discussions

No Known Activations