INDEX

Explanations

mentions of denial or being in denial

oai_token-act-pair · gpt-3.5-turbo

instances of the word "denial"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

IU

-0.93

liam

-0.85

orsi

-0.79

ulpt

-0.73

imen

-0.71

enegger

-0.71

âĢ¢âĢ¢

-0.69

encers

-0.69

rim

-0.68

tiny

-0.67

POSITIVE LOGITS

 denial

1.56

 denying

1.04

 deny

0.98

naissance

0.91

 denies

0.88

 acknowledgement

0.84

 mitigation

0.83

etheless

0.83

den

0.78

 denied

0.78

Activations Density 0.005%

mentions of denial or being in denial

instances of the word "denial"

No Comments

No Known Activations