INDEX

Explanations

phrases related to a daily dispatch newsletter

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "Dispatch" and related terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

egal

-0.78

umber

-0.78

urt

-0.73

rd

-0.73

gio

-0.69

--+

-0.68

unal

-0.67

ipedia

-0.66

arson

-0.64

atever

-0.64

POSITIVE LOGITS

 Dispatch

1.06

Dispatch

0.99

 dispatch

0.88

ables

0.86

DragonMagazine

0.79

Queue

0.78

hower

0.75

ivery

0.73

ãĥ¼ãĥĨãĤ£

0.73

able

0.73

Activations Density 0.018%

phrases related to a daily dispatch newsletter

occurrences of the word "Dispatch" and related terms

No Comments

No Known Activations