INDEX

Explanations

direct quotations or reported speech

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

estern

-0.75

anu

-0.75

ãĥ´

-0.67

acting

-0.65

plex

-0.64

entimes

-0.64

ãĥĩ

-0.63

 Unch

-0.62

mental

-0.62

ãĥ¼ãĥĨ

-0.60

POSITIVE LOGITS

 "...

0.97

 "âĢ¦

0.90

:"

0.90

"[

0.87

""

0.84

"'

0.82

"#

0.81

0.78

".

0.78

"(

0.77

Activations Density 0.116%

direct quotations or reported speech

No Comments

No Known Activations