INDEX

Explanations

informational content such as articles, documents, and transcripts for further reading

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Ã±o

-0.65

roit

-0.63

engeance

-0.61

xon

-0.61

pload

-0.61

charg

-0.60

ascal

-0.60

oso

-0.58

akedown

-0.58

uxe

-0.57

POSITIVE LOGITS

 aloud

1.50

just

1.15

 comprehension

1.10

dress

0.89

ied

0.88

 excerpts

0.85

mitt

0.84

 scriptures

0.81

mission

0.79

Write

0.78

Activations Density 1.536%

informational content such as articles, documents, and transcripts for further reading

No Comments

No Known Activations