INDEX

Explanations

phrases related to reputation or emerging as a significant entity

oai_token-act-pair · gpt-3.5-turbo

phrases indicating reputation or status

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

OTAL

-0.71

romy

-0.69

ength

-0.67

rost

-0.66

raq

-0.65

ongyang

-0.64

rent

-0.64

tails

-0.64

rone

-0.63

cloth

-0.62

POSITIVE LOGITS

pires

0.95

 synonymous

0.90

 well

0.89

pired

0.89

 embod

0.82

 unbeat

0.81

 invincible

0.80

 viable

0.78

 opposed

0.77

 adept

0.77

Activations Density 0.162%

phrases related to reputation or emerging as a significant entity

phrases indicating reputation or status

No Comments

No Known Activations