INDEX

Explanations

legal and bureaucratic language related to rules, instructions, and conditions

oai_token-act-pair · gpt-3.5-turbo

specific formal language and legal terminology

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

 gigs

-0.78

elight

-0.69

ï¸ı

-0.68

 vets

-0.68

 goodbye

-0.66

ricks

-0.65

 wisely

-0.64

everything

-0.64

 something

-0.64

 nicely

-0.63

POSITIVE LOGITS

 foregoing

1.55

 licensee

1.21

 applicant

1.20

 afore

1.17

 appellant

1.16

 Administrator

1.11

 petitioner

1.10

 respondent

1.08

 Applicant

1.07

 Secretary

1.06

Activations Density 0.399%

legal and bureaucratic language related to rules, instructions, and conditions

specific formal language and legal terminology

No Comments

No Known Activations