INDEX

Explanations

phrases emphasizing suitability, idealness, or desirability of subjects in various contexts

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

pronouns followed by descriptions

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

 puisque

-0.40

 porque

-0.40

 Però

-0.38

 puisqu

-0.36

 plazo

-0.35

 perquè

-0.33

parseColor

-0.33

 sueldo

-0.33

 morceaux

-0.31

Américaine

-0.31

POSITIVE LOGITS

 highly

0.81

 capable

0.67

highly

0.66

 arguably

0.65

 becoming

0.63

 potentially

0.63

 relatively

0.62

 extremely

0.61

 generally

0.60

 considerably

0.60

Activations Density 0.024%

phrases emphasizing suitability, idealness, or desirability of subjects in various contexts

pronouns followed by descriptions

No Comments

No Known Activations