INDEX

Explanations

phrases inviting interaction or communication

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.10.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

Mub

-0.61

 Parenthood

-0.61

rama

-0.60

 Phelps

-0.60

abama

-0.58

Boh

-0.57

 msec

-0.57

Maw

-0.57

 notch

-0.57

 Boko

-0.57

POSITIVE LOGITS

 free

1.05

free

0.93

FREE

0.91

 comfortable

0.88

 obligated

0.87

safe

0.86

 compelled

0.86

 confident

0.85

good

0.84

 FREE

0.81

Activations Density 0.028%

phrases inviting interaction or communication

No Comments

No Known Activations