INDEX

Explanations

terms related to interactions and their effects

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

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-res/layer_20/width_131k/average_l0_114

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

IFORN

-0.51

BibitemShut

-0.48

 newList

-0.47

uteen

-0.45

findpost

-0.45

ßerdem

-0.44

oubliez

-0.44

 Мексичка

-0.44

ähteet

-0.44

 Keuangan

-0.44

POSITIVE LOGITS

 interaction

1.21

 Interaction

1.12

interaction

1.10

 interactions

1.08

Interaction

1.04

 Interact

1.02

 Interactions

1.01

interactions

0.93

Interactions

0.93

interact

0.92

Activations Density 0.022%

terms related to interactions and their effects

No Comments

No Known Activations