INDEX

Explanations

proper nouns or names within a list

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate the presence of other entities or items in a list

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

govtrack

-0.71

fit

-0.61

 glim

-0.61

rison

-0.58

 owed

-0.55

 Procedure

-0.55

 Correct

-0.54

 culp

-0.54

oshenko

-0.53

agon

-0.52

POSITIVE LOGITS

 others

1.18

 countless

1.13

 innumerable

1.02

 myriad

1.01

 many

0.98

 dozens

0.93

 numerous

0.87

 else

0.83

 much

0.82

etc

0.81

Activations Density 0.168%

proper nouns or names within a list

phrases that indicate the presence of other entities or items in a list

No Comments

No Known Activations