INDEX

Explanations

text related to patents and legal documents

oai_token-act-pair · gpt-3.5-turbo

references to patents and patent applications

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

 Philly

-0.84

 Woody

-0.83

 goofy

-0.80

 moms

-0.79

 kids

-0.78

 yell

-0.77

 Boone

-0.77

 Freak

-0.75

 folks

-0.75

 Oprah

-0.75

POSITIVE LOGITS

sequently

1.18

 Whilst

1.08

 envis

1.07

Furthermore

1.05

Therefore

1.05

Firstly

1.04

 Firstly

1.02

respective

1.00

 Accordingly

0.96

Specifically

0.95

Activations Density 0.838%

text related to patents and legal documents

references to patents and patent applications

No Comments

No Known Activations