INDEX

Explanations

proper nouns or named entities

oai_token-act-pair · gpt-3.5-turbo

instances of numerical data or dates

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

 destro

-0.86

hement

-0.80

 helicop

-0.73

 Jagu

-0.72

nesday

-0.69

undai

-0.67

 favoured

-0.65

 acknow

-0.63

 abolish

-0.63

owicz

-0.62

POSITIVE LOGITS

Enlarge

0.83

Legendary

0.80

Overview

0.80

Temperature

0.76

PRESS

0.75

Includes

0.74

Volume

0.74

Updated

0.73

versions

0.73

ALT

0.72

Activations Density 0.214%

proper nouns or named entities

instances of numerical data or dates

No Comments

No Known Activations