INDEX

Explanations

references to mathematical or physical concepts, particularly related to "forces" or "factors."

oai_token-act-pair · gpt-3.5-turbo

phrases that denote quantities or factors related to physical forces or systems

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

days

-0.75

ipeg

-0.66

ãģ®å®

-0.64

iversary

-0.60

 Topic

-0.60

 Buckingham

-0.60

 Gazette

-0.59

 Idea

-0.59

 Hemp

-0.59

-0.58

POSITIVE LOGITS

 causation

0.78

upper

0.73

ocus

0.72

udic

0.66

ãĥ¼ãĥĨ

0.66

 mater

0.65

 exerted

0.65

effic

0.65

ativity

0.64

omaly

0.63

Activations Density 0.719%

references to mathematical or physical concepts, particularly related to "forces" or "factors."

phrases that denote quantities or factors related to physical forces or systems

No Comments

No Known Activations