INDEX

Explanations

words or acronyms related to the term "HS."

oai_token-act-pair · gpt-3.5-turbo

references to the acronym "HS," likely indicating a focus on healthcare or health services topics

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

bidden

-0.78

 Belg

-0.72

merce

-0.71

eering

-0.66

nesota

-0.66

 jour

-0.63

circle

-0.62

 Dane

-0.61

runtime

-0.60

 Samoa

-0.59

POSITIVE LOGITS

HS

1.51

Ds

1.06

HT

0.91

UFF

0.90

OOL

0.90

ustain

0.87

gow

0.85

IFT

0.85

UF

0.85

YS

0.84

Activations Density 0.007%

words or acronyms related to the term "HS."

references to the acronym "HS," likely indicating a focus on healthcare or health services topics

No Comments

No Known Activations