INDEX

Explanations

code structures and programming syntax

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 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

 noDo

-0.57

 geſch

-0.55

 cession

-0.55

vif

-0.54

 queſta

-0.53

ðsíða

-0.52

 pinulongan

-0.51

 zuſammen

-0.51

 pyridine

-0.50

asteroide

-0.50

POSITIVE LOGITS

 else

0.59

else

0.57

 otherwise

0.54

 yoksa

0.53

or

0.52

elif

0.51

 Otherwise

0.47

 sebaliknya

0.47

Otherwise

0.46

if

0.45

Activations Density 0.182%

code structures and programming syntax

No Comments

No Known Activations