INDEX

Explanations

names of specific entities, such as companies, people, or products

oai_token-act-pair · gpt-3.5-turbo

references to organizations, systems, or structures in a given context

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ABE

-0.64

&&

-0.63

 ensued

-0.60

 Chocobo

-0.60

!.

-0.59

 awaits

-0.57

!".

-0.57

lat

-0.56

cms

-0.55

otted

-0.55

POSITIVE LOGITS

 differently

1.29

 favorably

1.22

as

0.98

 negatively

0.95

 positively

0.90

 skept

0.85

 broadly

0.76

 harshly

0.75

 unfairly

0.75

 primarily

0.74

Activations Density 0.216%

names of specific entities, such as companies, people, or products

references to organizations, systems, or structures in a given context

No Comments

No Known Activations