INDEX

Explanations

phrases related to physical objects and actions, especially involving items being hidden or stored in different places

oai_token-act-pair · gpt-3.5-turbo

references to physical items and articles

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

NetMessage

-0.42

\\\\\\\\\\\\\\\\

-0.41

rawdownloadcloneembedreportprint

-0.41

--+

-0.40

++++++++++++++++

-0.40

------------------------------------------------

-0.36

 sudden

-0.36

pmwiki

-0.36

Population

-0.35

yss

-0.34

POSITIVE LOGITS

afety

0.68

hops

0.63

hips

0.62

hip

0.62

poons

0.60

uits

0.57

chool

0.55

paces

0.55

omething

0.54

etter

0.52

Activations Density 2.063%

phrases related to physical objects and actions, especially involving items being hidden or stored in different places

references to physical items and articles

No Comments

No Known Activations