INDEX

Explanations

sequences of whitespace tokens

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

assert and with

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 31-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_31/width_131k/average_l0_109

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.31.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

]=>

-0.60

 informée

-0.45

 noDo

-0.43

 laſt

-0.39

:✨

-0.38

 leaſt

-0.38

 Majefty

-0.37

 CreateTagHelper

-0.37

ArrowToggle

-0.36

>{@

-0.36

POSITIVE LOGITS

期刊论文

0.50

BagLayout

0.48

новниш

0.44

 sampel

0.42

 simplifié

0.42

기에

0.41

autoreleasepool

0.40

 waived

0.40

írus

0.40

Activations Density 0.004%

sequences of whitespace tokens

assert and with

No Comments

No Known Activations