INDEX

Explanations

negations, particularly related to actions or behaviors that are not being done

oai_token-act-pair · gpt-3.5-turbo

instances of the word "not" or phrases indicating negation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

 Expansion

-0.70

stakes

-0.67

 Pros

-0.66

 Films

-0.65

eers

-0.64

 Companies

-0.64

 Circuit

-0.62

Kry

-0.61

 Basics

-0.61

 Tours

-0.61

POSITIVE LOGITS

icably

1.41

epad

1.19

icable

1.15

 necessarily

1.09

hin

1.07

ched

0.97

orious

0.93

ifying

0.91

ices

0.89

necess

0.88

Activations Density 0.176%

negations, particularly related to actions or behaviors that are not being done

instances of the word "not" or phrases indicating negation

No Comments

No Known Activations