INDEX

Explanations

phrases associated with legal and procedural discussions regarding rights and evaluations

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

followed by descriptions or explanations

np_acts-logits-general · gemini-2.0-flash

expressed or stated

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 22-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_22/width_16k/average_l0_72

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.22.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

 tủ

-0.46

GIVEREF

-0.45

териалы

-0.44

 PREV

-0.44

ellido

-0.44

ährend

-0.43

pyplot

-0.43

 Blanchard

-0.43

@"/

-0.43

oulou

-0.42

POSITIVE LOGITS

 phrasing

1.23

 worded

1.16

ph

1.11

 framed

1.09

 expressed

1.08

 wording

1.05

 framing

1.04

 express

1.03

 Express

1.01

expressed

1.00

Activations Density 0.670%

phrases associated with legal and procedural discussions regarding rights and evaluations

followed by descriptions or explanations

expressed or stated

No Comments

No Known Activations