INDEX

Explanations

words related to accidents, disasters, and general negative events

oai_token-act-pair · gpt-3.5-turbo

coordinating conjunctions and phrases signifying processes or actions

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

abs

-0.78

scribe

-0.72

rack

-0.70

cho

-0.68

ivot

-0.67

dding

-0.66

apest

-0.66

Think

-0.66

ãĥĥãĤ¯

-0.66

irting

-0.65

POSITIVE LOGITS

 resulted

1.47

 lasted

1.47

 ended

1.39

 remained

1.36

 became

1.32

 culminated

1.31

 stayed

1.23

 plummeted

1.22

was

1.21

 proceeded

1.21

Activations Density 0.265%

words related to accidents, disasters, and general negative events

coordinating conjunctions and phrases signifying processes or actions

No Known Activations