INDEX

Explanations

words related to medical conditions and locations

oai_token-act-pair · gpt-3.5-turbo

names of locations or entities that start with 'H'

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

taboola

-0.99

ãĥīãĥ©ãĤ´ãĥ³

-0.83

ioxide

-0.78

 Expend

-0.72

 Sorce

-0.71

ciating

-0.68

BILITIES

-0.68

 Cosponsors

-0.67

Cola

-0.66

-+-+

-0.65

POSITIVE LOGITS

inki

1.04

icultural

0.83

zik

0.81

idan

0.80

wed

0.80

awei

0.79

kamp

0.77

iday

0.76

igans

0.76

quarters

0.75

Activations Density 0.103%

words related to medical conditions and locations

names of locations or entities that start with 'H'

No Comments

No Known Activations