INDEX

Explanations

names typically related to people such as "Holt"

oai_token-act-pair · gpt-3.5-turbo

mentions of specific names, primarily focusing on individuals with the surname "Holt" and "Bolt."

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

omething

-0.88

izoph

-0.87

pires

-0.84

ilial

-0.82

WithNo

-0.79

ystem

-0.78

ourning

-0.78

 Theft

-0.76

uties

-0.73

uded

-0.73

POSITIVE LOGITS

zman

1.04

 Bolt

0.83

tering

0.80

kamp

0.77

rix

0.77

zeb

0.74

ãĥĥãĥĪ

0.74

tsy

0.73

ãĤ¦ãĤ¹

0.72

zer

0.71

Activations Density 0.030%

names typically related to people such as "Holt"

mentions of specific names, primarily focusing on individuals with the surname "Holt" and "Bolt."

No Comments

No Known Activations