INDEX

Explanations

specific locations, institutions, and scientific references

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

<h1>

-0.60

И

-0.44

<h5>

-0.44

 Eliza

-0.42

UpdatedBy

-0.42

어

-0.41

import

-0.41

India

-0.41

-0.40

#!/

-0.40

POSITIVE LOGITS

ValueStyle

0.74

WriteBarrier

0.57

TagMode

0.56

TemporalType

0.56

存于互联网档案馆

0.50

Parcelize

0.50

KommentareTeilen

0.50

IntoConstraints

0.49

GOTREF

0.49

Datuak

0.49

Activations Density 0.036%

specific locations, institutions, and scientific references

No Comments

No Known Activations