INDEX

Explanations

instances of the number "seven" followed by a noun or an adjective

oai_token-act-pair · gpt-3.5-turbo

occurrences of the number seven

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

agate

-0.62

ller

-0.62

matter

-0.61

srf

-0.61

senal

-0.60

Document

-0.60

plom

-0.59

 Palestin

-0.58

 administr

-0.58

Uri

-0.57

POSITIVE LOGITS

teenth

1.38

teen

1.37

eenth

1.29

een

1.25

 hundred

1.01

fold

0.97

Seven

0.94

fif

0.93

irteen

0.91

eteen

0.91

Activations Density 0.014%

instances of the number "seven" followed by a noun or an adjective

occurrences of the number seven

No Comments

No Known Activations