INDEX

Explanations

phrases related to taking action or making decisions

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

ndra

-0.66

avage

-0.64

selection

-0.63

 pockets

-0.62

âĢ¢âĢ¢âĢ¢âĢ¢

-0.61

atton

-0.60

inguished

-0.59

lished

-0.58

clusions

-0.58

Operation

-0.57

POSITIVE LOGITS

 easy

0.86

 seriously

0.85

 upon

0.80

alian

0.77

 easier

0.77

 Seriously

0.75

easy

0.75

 stride

0.75

 apart

0.74

 pains

0.73

Activations Density 0.046%

phrases related to taking action or making decisions

No Comments

No Known Activations