INDEX

Explanations

numbers and abbreviations used to describe nutritional information

oai_token-act-pair · gpt-3.5-turbo

references to body mass measurements or nutritional content

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

OPLE

-0.81

 delinquent

-0.68

MENTS

-0.67

RELATED

-0.64

OUS

-0.64

MENT

-0.64

ARI

-0.64

 outfield

-0.62

 Samoa

-0.60

 hazard

-0.60

POSITIVE LOGITS

hirt

1.15

ites

1.09

erver

1.09

leep

1.05

poons

1.04

poon

1.04

ocket

1.04

pace

1.02

ession

1.02

ettings

1.01

Activations Density 0.024%

numbers and abbreviations used to describe nutritional information

references to body mass measurements or nutritional content

No Comments

No Known Activations