INDEX

Explanations

proper nouns and names

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

 tremend

-0.79

 citiz

-0.76

ometimes

-0.72

 footing

-0.68

ensibly

-0.67

 ingred

-0.66

cuts

-0.66

 hoard

-0.65

 endings

-0.65

 ageing

-0.64

POSITIVE LOGITS

↵

0.76

 Juda

0.71

Lua

0.69

 Judah

0.69

ãĤ¿

0.66

aho

0.66

assian

0.65

acio

0.65

 Emanuel

0.64

azel

0.63

Activations Density 0.225%

proper nouns and names

No Comments

No Known Activations