INDEX

Explanations

medical-related terms, especially those related to studies and conditions

oai_token-act-pair · gpt-3.5-turbo

references to analysis and evidence in a formal context

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

needed

-0.72

inctions

-0.64

Attempts

-0.60

 challeng

-0.58

rosso

-0.58

blem

-0.58

iquette

-0.57

 requisite

-0.56

 antidote

-0.55

ignty

-0.55

POSITIVE LOGITS

 standpoint

1.62

 onwards

1.40

 vantage

1.21

 perspective

1.18

 onward

1.12

 sources

0.96

 outset

0.85

 viewpoint

0.83

ratch

0.82

 perspectives

0.78

Activations Density 0.820%

medical-related terms, especially those related to studies and conditions

references to analysis and evidence in a formal context

No Comments

No Known Activations