INDEX

Explanations

emphasized terms indicating significance or special consideration

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

especially important or true

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-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

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

 BoxFit

-0.43

ロリ

-0.41

 éco

-0.36

 quoi

-0.35

OOC

-0.34

Const

-0.33

Clear

-0.32

 entraîne

-0.32

 couverts

-0.32

Connected

-0.31

POSITIVE LOGITS

 المعيارى

0.63

 Paglinawan

0.63

 deose

0.61

httphttps

0.60

 Signalez

0.59

+#+#

0.59

 dezelve

0.59

PYX

0.57

 zijne

0.57

DockStyle

0.56

Activations Density 0.033%

emphasized terms indicating significance or special consideration

especially important or true

No Comments

No Known Activations