INDEX

Explanations

medical-related terms or references in the text

oai_token-act-pair · gpt-3.5-turbo

references to medical topics or terminology

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

xual

-1.12

uble

-0.87

gger

-0.84

unin

-0.81

yip

-0.77

arget

-0.77

 capped

-0.72

 Unch

-0.71

Sic

-0.71

egg

-0.69

POSITIVE LOGITS

 practition

1.09

 professionals

0.94

 examiner

0.94

 practitioner

0.94

 marijuana

0.93

 practitioners

0.92

 care

0.88

 sciences

0.87

 diagnosis

0.82

 profession

0.82

Activations Density 0.024%

medical-related terms or references in the text

references to medical topics or terminology

No Comments

No Known Activations