INDEX

Explanations

superlatives and extremes, particularly focusing on size or importance

oai_token-act-pair · gpt-3.5-turbo

phrases emphasizing the significance or magnitude of specific concepts or issues

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

renheit

-0.77

ipel

-0.75

FIL

-0.70

PET

-0.70

][

-0.68

mid

-0.67

each

-0.67

rub

-0.67

ulum

-0.65

Phase

-0.64

POSITIVE LOGITS

 reason

1.22

 downside

1.19

 distinguishing

1.18

 caveat

1.16

 thing

1.16

 drawback

1.15

 takeaway

1.14

 question

1.10

 difference

1.08

 notable

1.06

Activations Density 0.170%

superlatives and extremes, particularly focusing on size or importance

phrases emphasizing the significance or magnitude of specific concepts or issues

No Comments

No Known Activations