INDEX

Explanations

the pronoun "he" in various contexts

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

he performing actions

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

vspace

-0.38

tic

-0.34

 distinct

-0.33

in

-0.33

όμε

-0.32

 आम

-0.31

ंदु

-0.31

 ecco

-0.31

よくある

-0.31

 across

-0.30

POSITIVE LOGITS

He

1.10

he

1.00

 ſhe

0.96

He

0.96

she

0.96

they

0.96

She

0.93

THEY

0.92

she

0.91

They

0.91

Activations Density 0.040%

the pronoun "he" in various contexts

he performing actions

No Comments

No Known Activations