INDEX

Explanations

phrases related to gratitude and recognition

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_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

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

notes

-0.66

DOC

-0.59

 scept

-0.59

 intangible

-0.59

DOC

-0.59

 simultane

-0.58

çī

-0.58

plates

-0.56

Reports

-0.56

\<

-0.56

POSITIVE LOGITS

schild

0.83

 necessarily

0.82

 bother

0.77

 hesitated

0.75

 hesitate

0.70

ajor

0.70

oppers

0.69

vana

0.68

enance

0.67

 tolerated

0.66

Activations Density 0.144%

phrases related to gratitude and recognition

No Comments

No Known Activations