INDEX

Explanations

occurrences of the word "every."

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-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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

cstdio

-0.48

Erreferentziak

-0.48

 ſche

-0.43

 adiv

-0.40

siąż

-0.39

లాలు

-0.39

 createState

-0.39

 ſch

-0.39

 Jefus

-0.38

HomeAsUpEnabled

-0.38

POSITIVE LOGITS

󠁴

0.56

assic

0.54

adays

0.52

 sehari

0.51

addCriterion

0.51

avelength

0.50

extAlignment

0.50

 Analytic

0.49

 week

0.49

Jeografia

0.49

Activations Density 0.011%

occurrences of the word "every."

No Comments

No Known Activations