INDEX

Explanations

phrases related to news reporting

oai_token-act-pair · gpt-3.5-turbo

phrases indicating general observations or statements about groups or phenomena

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

theless

-0.75

 arrang

-0.73

âĵĺ

-0.72

Secondly

-0.71

Otherwise

-0.62

sic

-0.62

result

-0.61

ONSORED

-0.60

 thereof

-0.60

 dispers

-0.59

POSITIVE LOGITS

 awoke

0.80

 woke

0.74

advertisement

0.70

 rejoice

0.69

reetings

0.67

 Slate

0.66

 Updated

0.64

joice

0.63

eah

0.63

 Hearthstone

0.62

Activations Density 0.674%

phrases related to news reporting

phrases indicating general observations or statements about groups or phenomena

No Comments

No Known Activations