INDEX

Explanations

the word "are" followed by an adjective

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "re."

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

BuyableInstoreAndOnline

-0.74

 lapt

-0.71

ertodd

-0.68

 tracker

-0.63

owl

-0.61

 Pebble

-0.61

 Kern

-0.60

Scroll

-0.59

 Seah

-0.59

 simulator

-0.59

POSITIVE LOGITS

aper

1.01

apers

0.93

ngth

0.92

selves

0.90

leases

0.89

becca

0.83

versible

0.82

neys

0.80

acters

0.80

akin

0.78

Activations Density 0.043%

the word "are" followed by an adjective

occurrences of the word "re."

No Comments

No Known Activations