INDEX

Explanations

phrases related to medical emergencies and hospital visits

oai_token-act-pair · gpt-3.5-turbo

references to hospitalizations and medical treatment following incidents

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 rebell

-0.81

hops

-0.74

soDeliveryDate

-0.73

 petitions

-0.72

 glitches

-0.71

 delet

-0.71

 disobedience

-0.70

 boycot

-0.70

 infring

-0.69

å¥³

-0.69

POSITIVE LOGITS

 hospital

1.39

ospital

1.23

 Hospital

1.19

 hospitals

1.08

 paramedics

1.03

 Medical

1.00

hospital

0.97

 Advocate

0.96

 Bellev

0.92

 Vanderbilt

0.92

Activations Density 0.099%

phrases related to medical emergencies and hospital visits

references to hospitalizations and medical treatment following incidents

No Comments

No Known Activations