INDEX

Explanations

references to medical emergency situations

oai_token-act-pair · gpt-3.5-turbo

terminology related to medicine

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 Lauder

-0.72

chnology

-0.69

 wart

-0.66

 innocence

-0.66

 indign

-0.62

 Orwell

-0.61

ACK

-0.60

 Canaveral

-0.59

 Audit

-0.58

pora

-0.58

POSITIVE LOGITS

iterranean

1.30

itating

1.24

itations

1.21

itated

1.21

ulla

1.20

icated

1.15

icum

1.14

ication

1.13

allion

1.13

itational

1.12

Activations Density 0.024%

references to medical emergency situations

terminology related to medicine

No Comments

No Known Activations