INDEX

Explanations

phrases indicating encouragement or suggestion for others to follow a certain action

oai_token-act-pair · gpt-3.5-turbo

the phrase "do the same."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

rs

-0.70

edia

-0.70

hower

-0.67

ontent

-0.67

neau

-0.66

availability

-0.65

iffe

-0.65

 therefore

-0.64

outh

-0.63

chief

-0.62

POSITIVE LOGITS

 unthinkable

1.31

 same

1.26

same

1.10

 opposite

1.09

 math

1.09

 deed

1.03

 maths

1.02

 bidding

0.99

 homework

0.99

 groundwork

0.97

Activations Density 0.051%

phrases indicating encouragement or suggestion for others to follow a certain action

the phrase "do the same."

No Comments

No Known Activations