INDEX

Explanations

medically related terms and concepts, including drug names, surgical procedures, and medical conditions

oai_token-act-pair · gpt-3.5-turbo

terms related to extreme or controversial issues

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

nesota

-0.66

akeru

-0.65

URA

-0.64

 contrace

-0.61

lishes

-0.60

ecast

-0.58

icago

-0.58

atican

-0.58

achev

-0.56

cffff

-0.56

POSITIVE LOGITS

 jewel

0.59

 Shoals

0.55

pton

0.53

Rust

0.53

Redditor

0.52

wid

0.52

 knot

0.51

 widow

0.49

ugs

0.49

tnc

0.49

Activations Density 0.858%

medically related terms and concepts, including drug names, surgical procedures, and medical conditions

terms related to extreme or controversial issues

No Comments

No Known Activations