INDEX

Explanations

mentions of individuals in texts, specifically in the context of various activities or interactions

oai_token-act-pair · gpt-3.5-turbo

references to individuals or personal entities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

rawdownloadcloneembedreportprint

-0.70

ãĤ¦ãĤ¹

-0.69

tails

-0.67

arp

-0.66

 Destination

-0.64

 Shogun

-0.64

Bey

-0.60

 Mang

-0.60

Oracle

-0.60

PER

-0.59

POSITIVE LOGITS

who

1.20

hips

1.16

paces

1.06

 residing

1.06

 whose

1.01

 whom

1.00

 possessing

0.99

 aged

0.98

who

0.94

folk

0.91

Activations Density 0.075%

mentions of individuals in texts, specifically in the context of various activities or interactions

references to individuals or personal entities

No Comments

No Known Activations