INDEX

Explanations

ideas or points to extract from information presented

oai_token-act-pair · gpt-3.5-turbo

phrases related to key summaries or takeaways from reports and discussions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

atron

-0.67

 david

-0.66

eering

-0.65

eers

-0.65

 advertised

-0.64

 icing

-0.64

 Kamp

-0.63

etheless

-0.60

 Rouge

-0.58

 ambig

-0.58

POSITIVE LOGITS

aways

1.67

overs

1.24

away

1.16

over

1.01

offs

1.01

 advantage

0.90

AW

0.87

OVER

0.87

uchi

0.81

 heed

0.78

Activations Density 0.041%

ideas or points to extract from information presented

phrases related to key summaries or takeaways from reports and discussions

No Comments

No Known Activations