INDEX

Explanations

comparative relationships and positive qualities

oai_token-act-pair · gpt-3.5-turbo

evaluative language that reflects opinions about individual worth and capability

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

 separat

-0.72

 occurs

-0.71

 VIDEOS

-0.71

Seg

-0.66

 excludes

-0.66

 disparity

-0.66

Demand

-0.66

ARM

-0.66

VERTISEMENT

-0.64

 Conver

-0.64

POSITIVE LOGITS

 proud

1.05

 happiest

0.99

 lucky

0.93

 wiser

0.90

 happy

0.88

 aware

0.86

 laughing

0.85

 pleased

0.85

 interested

0.84

ointed

0.84

Activations Density 0.648%

comparative relationships and positive qualities

evaluative language that reflects opinions about individual worth and capability

No Known Activations