INDEX

Explanations

numeric codes or abbreviations denoting specific items or concepts

oai_token-act-pair · gpt-3.5-turbo

abbreviations or acronyms commonly used in titles or headings

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

imens

-0.75

 tremend

-0.75

 tert

-0.74

 Debor

-0.65

ADRA

-0.65

izoph

-0.64

 Haram

-0.64

 citiz

-0.63

 occas

-0.63

 rall

-0.62

POSITIVE LOGITS

ingly

1.04

ings

1.03

warm

0.95

grass

0.93

hound

0.92

bread

0.89

erton

0.88

bones

0.88

orse

0.87

nuts

0.87

Activations Density 0.177%

numeric codes or abbreviations denoting specific items or concepts

abbreviations or acronyms commonly used in titles or headings

No Comments

No Known Activations