INDEX

Explanations

phrases related to organizations or entities being the subject of a specific action

oai_token-act-pair · gpt-3.5-turbo

the word "which" frequently in context to describing entities or relationships

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 VIDEOS

-0.64

 comprehension

-0.63

 Passing

-0.62

sed

-0.62

 Passage

-0.61

oken

-0.59

 indefinite

-0.58

Definition

-0.58

 lifes

-0.57

mma

-0.57

POSITIVE LOGITS

 oversees

1.20

 owns

1.18

 operates

1.06

 specializes

1.05

 regulates

1.04

 employs

0.99

 oversaw

0.98

 invests

0.97

 publishes

0.90

 governs

0.90

Activations Density 0.102%

phrases related to organizations or entities being the subject of a specific action

the word "which" frequently in context to describing entities or relationships

No Comments

No Known Activations