INDEX

Explanations

phrases related to excuses or cover stories

oai_token-act-pair · gpt-3.5-turbo

references to justifications or excuses for actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

enge

-0.75

ike

-0.74

verbs

-0.74

ishable

-0.72

omer

-0.69

vals

-0.67

irc

-0.66

istor

-0.64

ffen

-0.63

opathy

-0.61

POSITIVE LOGITS

 pretext

1.15

ual

0.97

ãĥ¼ãĥĨãĤ£

0.95

ress

0.82

aneers

0.81

ãĥķãĤ©

0.78

ãĥĨ

0.77

ually

0.76

DragonMagazine

0.75

 excuse

0.74

Activations Density 0.021%

phrases related to excuses or cover stories

references to justifications or excuses for actions

No Comments

No Known Activations