INDEX

Explanations

phrases indicating the ability or possibility of something

oai_token-act-pair · gpt-3.5-turbo

repeated occurrences of the word "can."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

pires

-0.77

soever

-0.72

thing

-0.61

 ancestor

-0.60

 Advis

-0.59

 Critics

-0.59

 vacated

-0.58

breakers

-0.58

wise

-0.57

furt

-0.57

POSITIVE LOGITS

't

1.46

 imagine

1.24

 attest

1.18

 assure

1.16

 tell

1.03

 relate

0.99

 envision

0.99

 honestly

0.98

berra

0.98

 understand

0.96

Activations Density 0.077%

phrases indicating the ability or possibility of something

repeated occurrences of the word "can."

No Comments

No Known Activations