INDEX

Explanations

complex scientific or medical terminology and data references

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

:✨

-0.90

 resourceCulture

-0.89

 myſelf

-0.86

 Houſe

-0.83

ChildScrollView

-0.82

 ſche

-0.81

 ujednoznacz

-0.81

 queſta

-0.79

 ſta

-0.79

RegressionTest

-0.77

POSITIVE LOGITS

0.32

 last

0.29

 pair

0.29

 leksikon

0.28

Sav

0.27

 lack

0.27

ante

0.26

 הז

0.26

strptime

0.26

Activations Density 0.006%

complex scientific or medical terminology and data references

No Comments

No Known Activations