INDEX

Explanations

terms related to writing and coding processes

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

UserScript

-0.43

Bä

-0.39

 rigtig

-0.38

 رف

-0.37

Uma

-0.36

 stiamo

-0.36

vų

-0.36

 descar

-0.36

rogram

-0.35

Wir

-0.35

POSITIVE LOGITS

verwijspagina

0.82

featureID

0.68

脚注の使い方

0.66

WriteTagHelper

0.66

 ब्रेकडाउन

0.63

addCriterion

0.58

parsedMessage

0.57

 ligiloj

0.57

tagHelperRunner

0.56

 AssemblyTitle

0.56

Activations Density 0.047%

terms related to writing and coding processes

No Comments

No Known Activations