INDEX

Explanations

terms related to specific fields and categories, particularly in legal, health, and technical contexts

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

New Auto-Interp

Top Features by Cosine Similarity

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

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

expandindo

-0.52

inguém

-0.47

the

-0.45

 Gelegenheit

-0.43

 bagay

-0.43

 metálica

-0.41

 kyse

-0.40

an

-0.40

 humaine

-0.40

식

-0.39

POSITIVE LOGITS

 Catawiki

0.65

own

0.60

⤹

0.56

@"/

0.52

UserScript

0.47

 utafitiHapana

0.46

sodik

0.46

 &___

0.44

FDRE

0.44

izr

0.44

Activations Density 0.717%

terms related to specific fields and categories, particularly in legal, health, and technical contexts

No Comments

No Known Activations