INDEX

Explanations

the word "be" followed by various descriptions or scenarios

oai_token-act-pair · gpt-3.5-turbo

phrases that express potentiality or possibilities

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Clear

-0.64

ware

-0.64

plex

-0.63

 tends

-0.61

 consumes

-0.60

often

-0.59

Kub

-0.59

Clear

-0.58

 assumes

-0.58

 seems

-0.57

POSITIVE LOGITS

ħĭ

1.02

 feas

0.85

 conce

0.77

 mistaken

0.77

 someday

0.76

ivably

0.75

efe

0.73

ebus

0.73

 possibly

0.69

0.68

Activations Density 0.190%

the word "be" followed by various descriptions or scenarios

phrases that express potentiality or possibilities

No Comments

No Known Activations