INDEX

Explanations

proper nouns related to individuals and their titles or positions in government

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_20/width_16k/average_l0_68

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

 ddelweddau

-0.52

ographia

-0.46

 Winaray

-0.45

Попис

-0.42

 optique

-0.42

ssf

-0.41

 متعلقه

-0.41

 óptica

-0.41

ographies

-0.40

MessageTagHelper

-0.40

POSITIVE LOGITS

 nakalista

0.54

 Deputy

0.46

 presided

0.44

 elected

0.43

 said

0.42

BeginInit

0.41

 deputy

0.40

μφωνα

0.40

 briefed

0.39

しゃれ

0.39

Activations Density 0.111%

proper nouns related to individuals and their titles or positions in government

No Comments

No Known Activations