INDEX

Explanations

references to legal or regulatory matters, focusing on the concept of laws

oai_token-act-pair · gpt-3.5-turbo

references to legal regulations or laws

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Dak

-0.79

atha

-0.73

 Mous

-0.68

 Crescent

-0.67

Mul

-0.65

 Objective

-0.65

alin

-0.64

spect

-0.61

 Ital

-0.60

sg

-0.59

POSITIVE LOGITS

 laws

1.17

 enforcement

1.16

 Laws

1.04

ĸļ

0.97

 governing

0.93

 lawy

0.92

enforcement

0.91

laws

0.90

iard

0.90

 prohibiting

0.89

Activations Density 0.015%

references to legal or regulatory matters, focusing on the concept of laws

references to legal regulations or laws

No Comments

No Known Activations