INDEX

Explanations

phrases indicating various experiences and actions

oai_token-act-pair · gpt-3.5-turbo

instances of having accomplished actions or ongoing states

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

 Apart

-0.73

 monog

-0.66

 Friendship

-0.63

 Outer

-0.59

aments

-0.59

TG

-0.59

 Ascension

-0.58

 Useful

-0.57

 doom

-0.56

 hostage

-0.55

POSITIVE LOGITS

 been

1.36

been

1.22

 become

1.14

 begun

1.12

 undergone

1.07

 gotten

1.03

 risen

1.02

 Been

0.93

 emerged

0.93

 gone

0.92

Activations Density 0.279%

phrases indicating various experiences and actions

instances of having accomplished actions or ongoing states

No Comments

No Known Activations