INDEX

Explanations

phrases related to reactions and responses

oai_token-act-pair · gpt-3.5-turbo

phrases indicating reactions or responses to announcements or events

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

Not in Any Lists

No Comments

Negative Logits

 entrusted

-0.78

agate

-0.64

hart

-0.62

ouf

-0.62

ciples

-0.61

yrights

-0.60

 Parameters

-0.60

dimension

-0.60

emy

-0.60

kun

-0.59

POSITIVE LOGITS

 disbelief

1.36

 dismay

1.35

 incred

1.34

 cheers

1.33

 applause

1.30

 skepticism

1.28

 silence

1.28

 laughter

1.25

 puzz

1.25

 indifference

1.25

Activations Density 0.213%

phrases related to reactions and responses

phrases indicating reactions or responses to announcements or events

No Comments

No Known Activations