INDEX

Explanations

specific patterns related to letter combinations and abbreviations

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

Two-letter combinations

np_acts-logits-general · gemini-2.0-flash

k, ck, pk, ik

np_acts-logits-general · gemini-2.5-flash-lite

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

izon

-0.55

awaiter

-0.51

VOA

-0.49

stahl

-0.49

慨

-0.48

tableFuture

-0.48

CEA

-0.47

esModule

-0.46

”

-0.46

useState

-0.46

POSITIVE LOGITS

0.75

kk

0.69

ambién

0.66

ke

0.63

ck

0.63

ky

0.62

kkkk

0.61

 Inscrivez

0.61

pk

0.61

ik

0.61

Activations Density 0.819%

specific patterns related to letter combinations and abbreviations

Two-letter combinations

k, ck, pk, ik

No Comments

No Known Activations