INDEX

Explanations

phrases related to detailed scientific or technical information

oai_token-act-pair · gpt-3.5-turbo

specific identifying features or characteristics related to substances and classifications

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

ndum

-0.65

 Canaver

-0.64

ffe

-0.63

emic

-0.59

 therap

-0.59

ppy

-0.59

 Moines

-0.58

 dere

-0.58

 mercy

-0.58

andum

-0.58

POSITIVE LOGITS

including

1.17

 namely

0.99

typically

0.94

particularly

0.90

which

0.88

containing

0.87

called

0.86

 including

0.85

available

0.84

often

0.83

Activations Density 0.302%

phrases related to detailed scientific or technical information

specific identifying features or characteristics related to substances and classifications

No Comments

No Known Activations