INDEX

Explanations

words related to medical conditions and treatments, particularly those ending in 'ed'

oai_token-act-pair · gpt-3.5-turbo

words related to concepts that are often marked by the suffix 'ed'

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

Story

-0.68

Nun

-0.65

 Bayer

-0.64

 Kiss

-0.63

ATT

-0.62

SIGN

-0.61

 sacrifices

-0.61

idine

-0.60

ãĥ¯ãĥ³

-0.60

 Hunters

-0.59

POSITIVE LOGITS

ifice

1.20

ed

1.03

ging

0.99

monton

0.97

dy

0.94

ocument

0.93

irect

0.92

ict

0.92

uction

0.89

gew

0.89

Activations Density 0.007%

words related to medical conditions and treatments, particularly those ending in 'ed'

words related to concepts that are often marked by the suffix 'ed'

No Comments

No Known Activations