INDEX

Explanations

phrases related to denial and rejection

oai_token-act-pair · gpt-3.5-turbo

phrases related to rejections and demands of information or claims

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

 Flickr

-0.71

 ILCS

-0.68

 Veter

-0.68

 Recover

-0.65

inki

-0.65

 Surv

-0.64

bucks

-0.63

Interstitial

-0.61

 Screw

-0.61

phabet

-0.59

POSITIVE LOGITS

 outright

1.17

 altogether

1.16

 notion

0.93

 suggestion

0.91

 entirely

0.85

 assertion

0.84

 advances

0.83

 accusation

0.80

 assertions

0.80

 accusations

0.78

Activations Density 0.340%

phrases related to denial and rejection

phrases related to rejections and demands of information or claims

No Comments

No Known Activations