INDEX

Explanations

phrases related to written communication

oai_token-act-pair · gpt-3.5-turbo

markers of sentence structure or formatting cues

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 frontline

-0.55

aky

-0.53

ussy

-0.52

emaker

-0.51

 proble

-0.51

 ÂŃ

-0.49

coni

-0.49

chuk

-0.49

afer

-0.48

ady

-0.47

POSITIVE LOGITS

huh

0.85

please

0.82

 please

0.82

etc

0.80

govtrack

0.79

channelAvailability

0.70

 whence

0.66

rete

0.65

 wherein

0.64

 yeah

0.63

Activations Density 0.235%

phrases related to written communication

markers of sentence structure or formatting cues

No Comments

No Known Activations