INDEX

Explanations

common phrases or connecting words that appear across different texts

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

aro

-0.71

uty

-0.69

ubs

-0.67

æ©

-0.66

ocaust

-0.66

oeuv

-0.66

afety

-0.64

WAYS

-0.64

enger

-0.63

onica

-0.63

POSITIVE LOGITS

we

0.75

rely

0.73

 pesky

0.71

 settles

0.66

adays

0.66

 wraps

0.62

ched

0.62

 they

0.61

he

0.59

 sounds

0.59

Activations Density 0.038%

common phrases or connecting words that appear across different texts

No Comments

No Known Activations