INDEX

Explanations

phrases related to author's notes or disclaimers

oai_token-act-pair · gpt-3.5-turbo

elements of disclaimers or notices commonly found in opinion pieces or articles

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 shutting

-0.90

 clutch

-0.78

 denying

-0.76

 manif

-0.74

 stopping

-0.72

 gunned

-0.71

 shut

-0.71

 knocking

-0.70

 shutdown

-0.69

 slowing

-0.69

POSITIVE LOGITS

 Please

1.22

 PLEASE

0.97

 Contains

0.96

 Unless

0.96

 Readers

0.95

Disclaimer

0.94

 Feel

0.94

 Therefore

0.93

 Published

0.88

Its

0.88

Activations Density 0.288%

phrases related to author's notes or disclaimers

elements of disclaimers or notices commonly found in opinion pieces or articles

No Comments

No Known Activations