INDEX

Explanations

words related to names or titles followed by quotation marks

oai_token-act-pair · gpt-3.5-turbo

instances of the word "as," indicating comparisons or relationships

oai_token-act-pair · gpt-4o-mini Triggered by @bot

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

âĸº

-0.68

ATIONS

-0.68

LESS

-0.68

+.

-0.65

###

-0.63

aster

-0.62

AGES

-0.60

BACK

-0.59

rise

-0.59

TOP

-0.59

POSITIVE LOGITS

 opposed

1.18

pires

1.10

phalt

1.04

pired

1.04

bestos

0.99

ide

0.97

 evidenced

0.97

 usual

0.94

piration

0.92

semb

0.90

Activations Density 0.057%

words related to names or titles followed by quotation marks

instances of the word "as," indicating comparisons or relationships

No Comments

No Known Activations