INDEX

Explanations

words related to regulation, specifically those involving the term "reg"

oai_token-act-pair · gpt-3.5-turbo

references to regulatory and governance-related topics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

lihood

-0.77

¿½

-0.75

 Dangerous

-0.73

 Misty

-0.73

 Pegasus

-0.67

 Brist

-0.67

 Benghazi

-0.67

Cue

-0.66

 Kinnikuman

-0.66

\\\\\\\\

-0.66

POSITIVE LOGITS

reg

1.08

ardless

1.07

inding

1.05

nant

1.02

istration

0.98

archs

0.95

aud

0.93

ritional

0.93

ener

0.92

ulators

0.91

Activations Density 0.005%

words related to regulation, specifically those involving the term "reg"

references to regulatory and governance-related topics

No Comments

No Known Activations