INDEX

Explanations

instances of the word "rules" with an activation level of 9 or 10

oai_token-act-pair · gpt-3.5-turbo

references to regulations or guidelines

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

ience

-0.86

 experien

-0.77

itate

-0.72

aged

-0.70

issance

-0.68

ONSORED

-0.68

 Bras

-0.68

apest

-0.67

ienced

-0.66

 Gothic

-0.63

POSITIVE LOGITS

 governing

1.07

 lawy

0.97

 Enforcement

0.96

breakers

0.92

book

0.90

 enforcement

0.90

 rules

0.88

 violations

0.87

breaker

0.84

books

0.83

Activations Density 0.024%

instances of the word "rules" with an activation level of 9 or 10

references to regulations or guidelines

No Comments

No Known Activations