INDEX

Explanations

phrases related to things being done in an aesthetically pleasing or well-executed manner

oai_token-act-pair · gpt-3.5-turbo

positive descriptions of qualities or characteristics

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 necessity

-0.69

hao

-0.66

resy

-0.66

 Ancients

-0.64

eway

-0.64

emic

-0.63

orem

-0.63

eport

-0.63

 restraints

-0.62

JV

-0.62

POSITIVE LOGITS

baum

0.98

 suited

0.95

 illustrated

0.81

 illustrates

0.79

istered

0.77

 Gleaming

0.76

 situated

0.76

ãĤ¼ãĤ¦ãĤ¹

0.74

 adapted

0.73

 tuned

0.71

Activations Density 0.043%

phrases related to things being done in an aesthetically pleasing or well-executed manner

positive descriptions of qualities or characteristics

No Comments

No Known Activations