INDEX

Explanations

words related to turning or events that involve a change or twist

oai_token-act-pair · gpt-3.5-turbo

references to turns or directional changes

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

ropolitan

-0.75

inately

-0.70

è¦ļéĨĴ

-0.67

aimon

-0.66

anyahu

-0.65

capacity

-0.64

 Mellon

-0.63

gdala

-0.62

llah

-0.62

 franc

-0.62

POSITIVE LOGITS

buck

0.96

about

0.95

around

0.85

abouts

0.84

coat

0.82

bull

0.80

overs

0.79

ips

0.79

wheel

0.78

over

0.77

Activations Density 0.035%

words related to turning or events that involve a change or twist

references to turns or directional changes

No Comments

No Known Activations