INDEX

Explanations

phrases related to identifying, addressing, or dealing with specific issues

oai_token-act-pair · gpt-3.5-turbo

coordinating conjunctions, particularly the word "and."

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

ESE

-0.71

 Broncos

-0.68

 Loren

-0.66

 Doodle

-0.66

izards

-0.64

éŃĶ

-0.63

 Dolphins

-0.63

olver

-0.62

 Patriots

-0.62

lon

-0.61

POSITIVE LOGITS

rogens

1.08

rogen

1.06

rew

0.94

 therefore

0.84

 consequently

0.78

 possibly

0.77

 distribute

0.74

0.72

 redistributed

0.72

romeda

0.70

Activations Density 0.253%

phrases related to identifying, addressing, or dealing with specific issues

coordinating conjunctions, particularly the word "and."

No Comments

No Known Activations