INDEX

Explanations

references to acts of violence or crime

oai_token-act-pair · gpt-3.5-turbo

conjunctions implying escalation or continuation of events

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Actions

-0.75

notations

-0.71

IST

-0.65

 Loss

-0.65

 Achievements

-0.63

Plus

-0.63

 Contributions

-0.63

 Improve

-0.63

ilty

-0.63

Color

-0.62

POSITIVE LOGITS

 began

1.24

 proceeded

1.22

 demanded

1.15

 commenced

1.09

 attempted

1.04

 gazed

1.00

 promptly

1.00

 waited

0.99

 begged

0.99

 confronted

0.98

Activations Density 0.126%

references to acts of violence or crime

conjunctions implying escalation or continuation of events

No Comments

No Known Activations