INDEX

Explanations

information related to educational topics, specifically microlearning and meditation practices

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs768-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

768

Data Type

torch.float32

Hook Point

blocks.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

edin

-0.78

aretz

-0.75

Jar

-0.65

EStream

-0.64

itals

-0.63

ĪĴ

-0.62

thora

-0.62

 outweigh

-0.60

 Hurt

-0.60

Lago

-0.60

POSITIVE LOGITS

 acronym

1.32

 abbre

1.28

 loosely

1.16

 shorthand

1.12

quartered

1.02

 coined

1.00

 slang

0.98

formerly

0.94

 comprised

0.93

 combining

0.92

Activations Density 7.056%

information related to educational topics, specifically microlearning and meditation practices

No Comments

No Known Activations