INDEX

Explanations

phrases related to the truthfulness or accuracy of statements

oai_token-act-pair · gpt-3.5-turbo

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ngth

-0.73

ividual

-0.73

 hail

-0.69

 Scores

-0.65

 freezes

-0.65

 clone

-0.62

owe

-0.61

 blasts

-0.60

anches

-0.59

icks

-0.59

POSITIVE LOGITS

 happening

1.39

true

1.37

 true

1.34

 untrue

1.21

false

1.11

 TRUE

1.08

 happen

1.04

 occurring

0.99

 happened

0.94

 achievable

0.91

Activations Density 0.216%

phrases related to the truthfulness or accuracy of statements

No Comments

No Known Activations