INDEX

Explanations

terms related to gears and gear changes

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

gear systems

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

 PropTypes

-0.52

 Pollock

-0.49

 Noyes

-0.48

 Macdonald

-0.48

 Abingdon

-0.47

 Colchester

-0.47

 Masson

-0.47

 Popp

-0.46

 Malone

-0.45

 onPostExecute

-0.45

POSITIVE LOGITS

 Gear

1.16

Gear

1.14

 gear

1.13

gear

1.09

 GEAR

1.08

GEAR

0.99

 Gears

0.96

 gears

0.93

gears

0.91

 gearing

0.83

Activations Density 0.005%

terms related to gears and gear changes

gear systems

No Comments

No Known Activations