INDEX

Explanations

combinations of shapes and numerical references related to actions or objects within a structured system

oai_token-act-pair · gpt-3.5-turbo

elements related to visual representation and organization

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Relief

-0.71

 Brotherhood

-0.65

Leaks

-0.63

 Canary

-0.61

 Jihad

-0.61

 Arab

-0.60

 Executive

-0.60

 Expedition

-0.59

 Nightmares

-0.59

 Lieberman

-0.59

POSITIVE LOGITS

mith

1.37

paces

1.29

pace

1.26

pring

1.22

hip

1.21

hips

1.11

ynthesis

1.10

cale

1.08

poons

1.03

peed

1.03

Activations Density 0.604%

combinations of shapes and numerical references related to actions or objects within a structured system

elements related to visual representation and organization

No Comments

No Known Activations