INDEX

Explanations

the word "in" and its variations in various contexts

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

built-in features

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

Amicalement

-0.69

 zoude

-0.65

ésult

-0.64

 dezelve

-0.61

RegressionTest

-0.60

Specifiche

-0.59

enezuela

-0.58

 parís

-0.57

 zijne

-0.57

pulseira

-0.57

POSITIVE LOGITS

builtin

0.79

 builtin

0.69

in

0.69

builtins

0.64

in

0.62

 built

0.60

 Built

0.59

built

0.58

Builtin

0.58

Built

0.57

Activations Density 0.002%

the word "in" and its variations in various contexts

built-in features

No Comments

No Known Activations