INDEX

Explanations

phrases related to natural disasters, particularly hurricanes

oai_token-act-pair · gpt-3.5-turbo

references to hurricanes and severe storm-related terminology

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

cius

-0.94

unal

-0.93

skirts

-0.86

UGC

-0.82

pires

-0.81

atum

-0.81

frame

-0.78

ĸļ

-0.77

yles

-0.76

orate

-0.76

POSITIVE LOGITS

 hurricanes

1.07

 landfall

1.06

 hurricane

1.05

 Irma

1.04

ricane

1.02

ricanes

1.01

 Katrina

1.00

 storms

0.95

 surge

0.92

 swarm

0.91

Activations Density 0.014%

phrases related to natural disasters, particularly hurricanes

references to hurricanes and severe storm-related terminology

No Comments

No Known Activations