INDEX

Explanations

the word "bound" followed by another word

oai_token-act-pair · gpt-3.5-turbo

repeated instances of the word "bound," indicating a focus on constraints or connections

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

issance

-0.99

amera

-0.72

ciation

-0.70

 Avalon

-0.67

 fielded

-0.66

HAEL

-0.63

odder

-0.63

soDeliveryDate

-0.62

blems

-0.61

asio

-0.60

POSITIVE LOGITS

lessly

1.15

less

1.14

ging

0.94

binding

0.94

lessness

0.93

fold

0.92

gling

0.91

pipe

0.86

post

0.84

sym

0.80

Activations Density 0.028%

the word "bound" followed by another word

repeated instances of the word "bound," indicating a focus on constraints or connections

No Comments

No Known Activations