INDEX

Explanations

the words "but", "also", and other similar contrasting or connecting terms

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

ouri

-0.73

orc

-0.72

apter

-0.70

coat

-0.70

irm

-0.68

dl

-0.67

 Refresh

-0.65

ct

-0.64

bot

-0.64

hill

-0.63

POSITIVE LOGITS

 also

1.18

 ALSO

1.08

also

0.90

 actively

0.86

 secondly

0.85

 moreover

0.84

 importantly

0.84

 furthermore

0.81

 downright

0.78

 actually

0.78

Activations Density 0.049%

the words "but", "also", and other similar contrasting or connecting terms

No Comments

No Known Activations