INDEX

Explanations

phrases starting with "And" with a higher weight on the ones where the next word has a relatively lower frequency in the document

oai_token-act-pair · gpt-3.5-turbo

phrases starting with "And" that indicate a continuation or elaboration of thoughts

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

Not in Any Lists

No Comments

Negative Logits

âĦ¢:

-0.79

 Â®

-0.77

mug

-0.65

 Droid

-0.60

!)

-0.60

 adorable

-0.60

friendly

-0.58

 paws

-0.57

paw

-0.57

worldly

-0.56

POSITIVE LOGITS

entimes

0.97

 yeah

0.85

 secondly

0.80

ertodd

0.78

 glim

0.76

yss

0.72

0.69

 luckily

0.69

Dialogue

0.68

hiba

0.67

Activations Density 0.212%

phrases starting with "And" with a higher weight on the ones where the next word has a relatively lower frequency in the document

phrases starting with "And" that indicate a continuation or elaboration of thoughts

No Comments

No Known Activations