INDEX

Explanations

the word 'te' or 'te' followed by a number

oai_token-act-pair · gpt-3.5-turbo

the word "te" in various contexts throughout the document

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

interrupted

-0.73

ipolar

-0.68

antha

-0.66

olicy

-0.62

 sweep

-0.62

lessly

-0.62

usher

-0.61

rha

-0.60

OST

-0.60

 allowances

-0.60

POSITIVE LOGITS

brate

1.54

brates

1.48

ller

1.20

llers

1.18

levision

1.18

achers

1.12

llo

1.09

lla

1.04

achable

1.04

eve

1.03

Activations Density 0.033%

the word 'te' or 'te' followed by a number

the word "te" in various contexts throughout the document

No Comments

No Known Activations