INDEX

Explanations

tweets related to retweeting

oai_token-act-pair · gpt-3.5-turbo

Twitter-related terms or references to tweets

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 scanners

-0.73

 scope

-0.72

 peripheral

-0.71

 auxiliary

-0.70

SAM

-0.69

Accessory

-0.68

 circum

-0.67

 motions

-0.66

 ropes

-0.65

 mechanics

-0.65

POSITIVE LOGITS

weet

4.81

Tweet

1.33

alted

1.23

 retweet

1.22

 tweet

1.12

sweet

1.11

 Tweet

1.08

ilver

1.05

pring

1.02

hirt

0.99

Activations Density 0.011%

tweets related to retweeting

Twitter-related terms or references to tweets

No Comments

No Known Activations