INDEX

Explanations

superlatives indicating the most extreme degree or ranking of something

oai_token-act-pair · gpt-3.5-turbo

instances of the word "most" used to emphasize extremes or significance

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

æ©

-0.74

pload

-0.70

heid

-0.67

oak

-0.66

rompt

-0.66

 Films

-0.65

olor

-0.64

halla

-0.63

instead

-0.63

 Travels

-0.63

POSITIVE LOGITS

 important

1.02

likely

0.93

 obvious

0.92

 profitable

0.91

 notable

0.90

 interesting

0.88

 prominent

0.88

 likely

0.88

 plausible

0.88

 exciting

0.87

Activations Density 0.063%

superlatives indicating the most extreme degree or ranking of something

instances of the word "most" used to emphasize extremes or significance

No Comments

No Known Activations