INDEX

Explanations

phrases related to denial or disclaimers

oai_token-act-pair · gpt-3.5-turbo

statements expressing denial or lack of involvement

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 beware

-0.70

utenberg

-0.68

surprisingly

-0.68

 hailed

-0.63

 srfAttach

-0.61

 plenty

-0.59

ismo

-0.59

 surprisingly

-0.59

 badass

-0.57

 hilarious

-0.57

POSITIVE LOGITS

nor

1.52

 whatsoever

1.44

 anybody

1.19

 anything

1.09

 anymore

1.04

nor

1.00

 âĢ¦"

0.98

any

0.96

0.95

 ..."

0.94

Activations Density 0.536%

phrases related to denial or disclaimers

statements expressing denial or lack of involvement

No Comments

No Known Activations